Tag Archives: AI

The Abomination of AI – part 8 – summary and recap

Posted on by

This final post recaps what we’ve learnt about the runaway nature of the AI industry, how it undermines free markets, and how we can make a difference. The core question is not what can AI do, but what should AI do?

This is the last of the series of blogs based on my keynote “The abomination of AI” at ICoSCI 2026.  Each has an accompanying segment of the video and slides from the talk as well as detailed notes and references.  Section numbers refer to the full report which will be released in the final blog.   The slide thumbnails in the text correspond to the slides in the navigation panel below.  The presentation can be played below, or opened full screen. The full length video, complete slides and further information can be found at: https://alandix.com/academic/talks/ICOSCI-2026-abomination-of-AI/

Previously …

§1.  Every industry is driven by profits and power, but there is something about the nature of AI itself, which interacts with the nature of market forces in the world that is problematic and is different from other technologies.

§2.  Can any technology be neutral?  AI can be used for good purposes, such as advances in healthcare.  It can also have bad outcomes such as bias in the criminal justice system or online exploitative pornography.  Perhaps most often it is creating the frivolous or even ugly.

§3.  The obvious impact of AI is in the things it does directly. Some technologies also change the very nature of society, affecting even those who do not use them. Cars are an obvious example.  AI is also such a technology.

§4.  Doomsayers worry about the point when AI becomes sentient, outgrowing its creators.  The real danger is more insidious: the massive financial and human impacts of AI seem almost obscene.

§5.   Network externalities, the way one person’s use of AI and digital tech changes its value for others, creates positive feedback loops, leading to runaway growth and emergent monopolies, the nemesis of free markets. This the very nature of digital technology and AI breaks free markets leading to runaway inequality, even with the best intentions of industry … but some tech companies further exploit these effects.

§6.   Runaway growth of AI is not painless – opportunity costs of investment and human costs of lost jobs.  Gains may be transitory – buy-now-pay-later tech risk tying users into spiralling costs.

§7.   It all seems too big, requiring national and international responses.  But we can make a difference using appropriately chosen small AI (including none). Plus, this good use of AI is good for business too.

8.  In summary

So, in summary,  AI can do amazing good things, but often also bad things.

More crucial is how AI is shaping society.  We have to think explicitly about this, because AI has its own dynamic.  That dynamic is not good by its nature, so we have to control it to make it serve society. Some of this needs action at a governmental and intergovernmental level, because this is so big compared to most countries.

However there are things you can do.  You can choose to sometimes not use AI or use small AI, but always to try to use AI appropriately rather than just throwing it at a problem and wiping your hands of the wider impact.

The core question is not to think solely about what can AI do?  While it still does some things frighteningly badly, AI is getting better and better at doing more and more.

But the big question, the Big, BIG question is what should AI do?

And that’s the question we need to ask ourselves continually both in our individual work and at societal level.

AI will be an abomination, but only if we let it be.

Updates

It is now four months since I gave the talk at ICoSCI 2026 on which this blog series has been based.  In that short time there have been many changes, some strengthening the arguments and some challenging them.   In addition, I’ve had helpful feedback from several people, especially extensive comments by Mark Bernstein; so many thanks to Mark and others who have engaged with this series.

I’ve written updates at the end of several blogs.  In part 3 “A different kind of apocalypse” recent reports of agentic AI ignoring guardrails make Terminator-style AI devastation, seem less distant.  Of course, since those updates the publicity around Claude Mythos’ ability to find bugs in established codebases meant that Anthropic deemed it too dangerous to release without allowing selected partners to use it first to check their own security [An26a].  While this may be in part a PR exercise, it is being taken seriously by government and pan-government organisations [AISI26,Go26].  Furthermore, as well as these external threats, Anthropic are also monitoring for the potential of AI ‘sabotage’, that is:

“when an AI model with access to powerful affordances within an organization uses its affordances to autonomously exploit, manipulate, or tamper with that organization’s systems or decision-making in a way that raises the risk of future catastrophic outcomes”  [An26b]

Updates at the end of part 6 “should we worry?“ reinforce the difficulty of switching AI models and the way OpenClaw has emphasised the under-pricing of AI use plans, and hence the way that these might adjust (upwards!) over time, just as we’ve seen with other forms of digital technology.  At the end of part 7 “what can we do?” there is a lovely example of really smart AI, combining AI and plain old computing to achieve better and cheaper outcomes.

In addition to these updates, recent developments (since the updates!) and comments have raised a couple of issues that I’d like to address.

Size matters

Since giving the talk in January, there have been further developments in reducing the costs of AI, not least DeepSeek’s V4 release, which has focused on making model training and execution more efficient [DS26].  Nvidia have released open source models designed to run on local small-scale (as in not more than a few $1000 Nvidia chips) installations [Ca26,Br26], some of these are designed for specialised applications, but some more general purpose; it could be that they are positioning themselves to spread their market beyond the small number of AI software mega-corporations, but in the process may weaken the emergent monopolies of these software players.  However Nvidia’s own near monopoly AI hardware position is being challenged by DeepSeek’s use of Huawei chips [CC26].

This seems to suggest several potential scenarios:

  1. The big players (OpenAI, Anthropic) see off the cheaper, but less powerful alternatives, maybe using market dominance to retain near monopoly positions as discussed in section 5 (blog part 4 and part 5)
  2. The lean, mean models become powerful enough to open up the market fully, so that the current mega companies and their investors lose their ‘bet’ on market dominance, leading to massive drops in their market values, and potentially a major stock exchange crash.
  3. The cheaper models become viable alternatives, but do not immediately compete on sheer power and corporate commitment leaving the mega AI corporations strong, but less all encompassing, and making the individual solution strategies in part 7 easier.

The impact of (2) on the global economy would be pretty disastrous, especially following the massive hits of US-Israel/Iran and Ukraine/Russia wars, so, on balance, the softer movement of scenario (3) feels like the best outcome.

Perversely, the big AI companies would be likely to weather the storm of (2) as the investment already committed provides a cash buffer. This was certainly the case during the dot-com period in which tech companies with second/third round investment before the crash often survived, including Lastminute.com, the IPO of which triggered the market re-evaluation of tech in 2000.  The founders and early funders will see paper devaluations, but otherwise still be in control of huge businesses; the smaller, more recent investors will lose however, including many global pension funds.

Cats and Consummate Consumerism

Section 3 (blog part 2) is, in part, quite dismissive of the vast volume of ‘frivolous’ use of generative AI.  Later, I hope that is clarified (e.g. the example of the Doctor’s Kitchen app) that this does not mean criticising all personal use of AI — appropriate use of AI in can be very beneficial, in particular allowing far more individualised access to digital technology.  Indeed, LLMs are already democratising access to many forms of professional advice that are beyond the reach of individuals and small businesses [Fu26].

However, that does leave the cats.   If that is what people want to create and view, surely that is their business?

In some ways these uses of AI are the ultimate form of consumerism — like the boxfuls of unused plastic toys, kitchen appliances that lie in the dark recess of cupboards, the 1.6 billion items of clothing in UK wardrobes that have never been worn [BBC22] — but now all digital, thrust before us by the relentless algorithms of social media.  Items we never knew we wanted instantly become essential, produced apparently for free and provided in precisely the quantity and kind that makes us want more.

Is this a choice, when the algorithms know how to nudge and channel us [HS26], when LLMs have learnt the lessons of the confidence trickster, and when the content itself is addictive [KK25]?  Is this a free market equivalent of Opium Wars?

For individuals many of the costs are effectively hidden, especially at the point of use.  Just as no fleece wearer or takeaway coffee drinker deliberately chooses to put microplastics in breast milk, the environmental and social impacts of digital and AI products are often physically and temporally distant and in many cases suffered by others [Ma24].

This distancing is in part due to digital communication and in part the diffuse relationship between the loci of production and use, especially when a large proportion of cost is in training.  However, the distancing is in part deliberate, not least the under-pricing of services to build reliance, a trick that has been part of digital products almost since their onset and very much in the playbook of the neighbourhood drug dealer.

One reason for listing the almost unbelievable facts and figures of AI growth in part 3 (§4,2) is to force us to face these choices explicitly.

The speed of change …

As is evident things are moving rapidly. This said although the details are changing, many of the large-scale impacts of AI on society and economics outlined in this series build on longer term trends in digital technology that have been evident at least since the turn of the Millennium.

With so many technologies in the past the societal impacts have only become apparent in hindsight.  With AI there are surprises, especially in terms of its spurts of almost unimaginably rapid progress, but also we are increasingly aware of the dangers and pitfalls.  The issues describe here are part of this conversation, aiming to ensure that we enter this exciting and dangerous time with eyes wide open.

Coming soon …

If you are interested in these issues, look out for the book AI or Social Justice, which Clara Crivellaro and I are currently working on.  The book website already includes a growing collection of resources including case studies and videos. .

References

[AISI26] AI Security Institute (2026) Our evaluation of Claude Mythos Preview’s cyber capabilities. AI Security Institute, Department of Science, Innovation and Technology. Apr 13, 2026. https://www.aisi.gov.uk/blog/our-evaluation-of-claude-mythos-previews-cyber-capabilities

[An26a]  Anthropic (2026).  Project Glasswing: Securing critical software for the AI era. Accessed 4th May 2026. https://www.anthropic.com/glasswing

[An26b]  Anthropic (2026).  Sabotage Risk Report: Claude Opus 4.6. Accessed 4th May 2026.   https://anthropic.com/claude-opus-4-6-risk-report

[BBC22]  BBC News (2022).  UK wardrobes stuffed with unworn clothes, study shows.  BBC News, 7 October 2022. https://www.bbc.co.uk/news/science-environment-63170952

[Br26]  Kari Briski (2026).  NVIDIA Launches Nemotron 3 Nano Omni Model, Unifying Vision, Audio and Language for up to 9x More Efficient AI Agents.  Nvidia blog.  April 28, 2026.  https://blogs.nvidia.com/blog/nemotron-3-nano-omni-multimodal-ai-agents/

[CC26]  Caiwei Chen (2026).  Three reasons why DeepSeek’s new model matters: The long-awaited V4 is more efficient and a win for Chinese chipmakers.  MIT Technology Review, April 24, 2026  https://www.technologyreview.com/2026/04/24/1136422/why-deepseeks-v4-matters/

[Ca26] Bryan Catanzaro (2026). NVIDIA Launches Open Models and Data to Accelerate AI Innovation Across Language, Biology and Robotics.  NVIDAI Blog, October 28, 2025.  https://blogs.nvidia.com/blog/open-models-data-ai/

[DS26]  DeepSeek-AI (2026).  DeepSeek-V4:Towards Highly Efficient Million-Token Context Intelligence.  Accessed 29th April 2026.  https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro/blob/main/DeepSeek_V4.pdf

[Fu26]  Maria Ines Fuenmayor (2026). The Privilege of Refusing AI. Codified, Medum,  Apr 22, 2026.  https://codifiedai.substack.com/p/the-privilege-of-refusing-ai

[Go26]  Gordon M. Goldstein (2026). Six Reasons Claude Mythos Is an Inflection Point for AI—and Global Security.  Council on Foreign Relations, April 15, 2026. https://www.cfr.org/articles/six-reasons-claude-mythos-is-an-inflection-point-for-ai-and-global-security

[HS26]  Kali Hays, Nardine Saad and Regan Morris, (2026).  Campaigners welcome Meta and YouTube’s defeat in landmark social media addiction trial.  BBC News, 25 March 2026.  https://www.bbc.co.uk/news/articles/c747x7gz249o

[KK25]  Kooli, Chokri, Youssef Kooli, and Eya Kooli (2025). Generative artificial intelligence addiction syndrome: A new behavioral disorder?.  Asian Journal of Psychiatry 107:104476.  https://doi.org/10.1016/j.ajp.2025.104476

[Ma24]  Murgia, Madhumita (2024). Code Dependent: How AI Is Changing Our Lives. Picador.

 

Facial recognition — what does accuracy mean?

Posted on by

A Guardian article at the weekend reported on the increasing number of people being ejected from stores after being misidentified by facial recognition systems as past shoplifters [Mu26].   This commercial use of facial regulation has even less oversight than police use, which has also been causing alarm. The people at the centre of the report were eventually offered gift vouchers by the shops concerned, but only after considerable personal embarrassment and lengthy and complex processes to clear their names (or to be precise faces).

According to the article Facewatch, the company providing the facial recognition service, claim a 99.98% accuracy rate.  This sounds high.  Does this mean that the cases reported are rare, albeit unfortunate, incidents?

Let’s unpack this a little.

According to the UK Office of National Statistics annual report on Crime in England and Wales, there are just over half a million cases of shoplifting a year  [ONS26]; the Facewatch web site offers a higher figure of 2 million across the whole UK, maybe attempting to take into account underreporting [FW26].  Let’s use this larger figure.

In the UK there are about 55 million adults, assuming on average of one shop visit per day, that is about 20 billion shopping visits per year.  So that means shoplifting accounts for just one visit in 10,000.1

So, if a facial recognition systems said no-one was a past shoplifter, it would attain 99.99% accuracy!2  If on the other hand the accuracy is equal for shoplifters and non-shoplifters (that is false positive and false negative rates are the same), then there would be one misidentified innocent for every correctly identified shoplifter — hardly rare.  If we use the ONS shoplifting figures, this rises to three misidentifications for each correct one.

One assumes that Facewatch adjusted the system recognition thresholds to have a lower false positive rate (wrongly accused) than this, instead accepting a greater proportion of missed true shoplifters, but in this case an overall 99.98% figure is unachievable.  Most likely the reported figure it is based on training data with, perhaps equal numbers of photos of shoplifters and non-shoplifters (essential to allow effective learning), so the 99.98% accuracy figure refers to this data not the numbers of each encountered in realistic (let alone real) use.

In both this case and others, such as rare disease diagnosis, seemingly high stated accuracy rates may not be as good as they at first seem, and certainly need a lot of context to be meaningful. As is clear this is by no means an abstract mathematical discussion, but one that affects real lives.  In the case of the use of facial recognition, the article also reminds us that these kinds of systems often have lower accuracy rates, and in particular higher false positive rates (that is wrongly accused) for black and asian people and for women in particular.

 

References

[FW26]   Facewath (2026).  Home page. Accessed 4th May 2026.  https://www.facewatch.co.uk

[Mu26]  Jessica Murray.  Guilty until proven innocent: shoppers falsely identified by facial recognition system struggle to clear their names.  The Guardian, 3 May 2026.  https://www.theguardian.com/technology/2026/may/03/guilty-until-proven-innocent-shoppers-falsely-identified-by-facial-recognition-struggle-to-clear-their-name

[ONS26]  Office of National Statistics (2026).  Crime in England and Wales: year ending December 2025.  ONS Centre for Crime and Justice, 23 April 2026.  https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/bulletins/ crimeinenglandandwales/yearendingdecember2025

 

  1. It is really hard to keep track of these huge numbers.  I’m expert at it, but I initially made a small slip and was out by a factor of 20.[back]
  2. When I read accuracy figures in academic papers on machine learning, I often do the equivalent calculation for a trivial classifier … as in this case, it is often no worse than the algorithm.[back]

The Abomination of AI – part 7 – what can we do?

Posted on by

It all seems too big, requiring national and international responses.  But we can make a difference using appropriately chosen small AI (including none). Plus, this good use of AI is good for business too.

This is the seventh of a series of blogs based on my keynote “The abomination of AI” at ICoSCI 2026.  Each has an accompanying segment of the video and slides from the talk as well as detailed notes and references.  Section numbers refer to the full report which will be released in the final blog.   The slide thumbnails in the text correspond to the slides in the navigation panel below.  The presentation can be played below, or opened full screen. The full length video, complete slides and further information can be found at: https://alandix.com/academic/talks/ICOSCI-2026-abomination-of-AI/

Previously …

§1.  Every industry is driven by profits and power, but there is something about the nature of AI itself, which interacts with the nature of market forces in the world that is problematic and is different from other technologies.

§2.  Can any technology be neutral?  AI can be used for good purposes, such as advances in healthcare.  It can also have bad outcomes such as bias in the criminal justice system or online exploitative pornography.  Perhaps most often it is creating the frivolous or even ugly.

§3.  The obvious impact of AI is in the things it does directly. Some technologies also change the very nature of society, affecting even those who do not use them. Cars are an obvious example.  AI is also such a technology.

§4.  Doomsayers worry about the point when AI becomes sentient, outgrowing its creators.  The real danger is more insidious: the massive financial and human impacts of AI seem almost obscene.

§5.   Network externalities, the way one person’s use of AI and digital tech changes its value for others, creates positive feedback loops, leading to runaway growth and emergent monopolies, the nemesis of free markets. This the very nature of digital technology and AI breaks free markets leading to runaway inequality, even with the best intentions of industry … but some tech companies further exploit these effects.

§6.   Runaway growth of AI is not painless – opportunity costs of investment and human costs of lost jobs.  Gains may be transitory – buy-now-pay-later tech risk tying users into spiralling costs.

 

7.  What can we do?

These issues all seem too big, frighteningly so.   So what can you do?

You might be a policy maker, or on a government committee that’s advising governments.  If so, you might be in position to make changes at that scale.  Most do not have such high-level influence, but there are changes you can make within your own spheres to help ameliorate some of these potential dangers.  I’ll focus on the UX designer or AI developer, but some of the ideas are ones you might be able to adopt in your own personal use or within an organisation.

 

7.1  No AI

One option is to simply, say “no” to AI.

If you are a designer, ask, “do I need AI at all in my project?”  Of course, everybody now expects every product to say ‘AI powered’, so you may not be able to avoid AI altogether, but it could be very simple AI.  But do ask whether you need it at all, if you don’t, why are you feeling you need to use it?

 

7.2  Small AI

If you do decide to use AI, you can opt for small AI.

If you are using language models or other generative AI, you might use smaller models, the kind that have been deliberately designed to be able to run on less powerful hardware. There are many good reasons to do this.  Indeed, Apple have been encouraging smaller AI because they want the AI to run on people’s personal devices, not just in the cloud.  This is because privacy is a strong part of the company brand.

Where it is appropriate you could use traditional AI, which is usually much smaller in terms of memory and computation.

Purely from a technical perspective, there are some really interesting research challenges in this area, both in terms of human computer interaction (see my 2024 talk on ‘Patient Interaction’ [Dx24]) and also pure technical AI.

Images: [Di22,Sa23,Dx25,DS25]

You’ll have seen some of the modifications of algorithms that are transforming this landscape including open LLaMa [ZD22,TH23], LORA [HS22] and LiGO [WP23].  DeepSeek [DS24,DS25] made waves when US export restrictions on NVIDIA chips forced Chinese innovators to adopt a far leaner and smarter approach to LLM development [LF24].  Debatably DeepSeek’s learning might have piggybacked off some of the other LLMs [We25b], but certainly at execution time it used far less resources than other LLMs at time. Now other LLMs have adopted lessons from DeepSeek, and all are looking to perform more efficiently, so there is small shift in thinking away from a simplistic ‘bigger is better’ approach [Hi20].

 

7.3  When to use AI

There is also a choice about when to use AI.

The most obvious use of AI is at execution time in a user interface or delivered application as part of the service provided.  This can of course be small AI or no AI at all even.

But you can also use AI at design time.  You might use big AI to create small AI for the delivered system, for example using techniques to compress the model.  You can also use AI as part of the UX process to critique a user interface, create rapid prototypes, or propose design ideas [Dx26b].  In addition, AI-based coding tools can create AI-free (or low-AI) systems.

Crucially, if you use big AI to help create a (smaller) product, it effectively gets reused again and again and again and again.   So it’s less expensive – both moneywise expensive to a company, but also less expensive in terms of its impact on the environment and society.

In fact, this is really powerful use of AI. For instance, one of the things I argue elsewhere is that AI critiques of UIs will be far better for accessibility than even the best designers.  This is in part because it is really hard for us to think about even obvious diversity such as what’s it like to be blind or deaf, or have a physical disability, an automated design tool can check a concept or prototype against vast numbers of different types of perceptual and physical abilities as well as combinations.  Even more important, it is almost impossible for us to imagine what it’s like to be somebody who thinks differently, for example somewhere distant from ourselves in a neurodivergent space .  I don’t think AI will be good at this, but I think it will be better than we are.

 

7.4  How to use AI

Finally, if you are using AI think carefully about the kind of AI you are going to use, and how to incorporate it into a system.  For many years I’ve talked about appropriate intelligence, most often in relation to AI error and the need to design human–AI systems that together are robust and effective, not focusing in the AI accuracy alone [DB00,BD23].  However, the same lesson can be applied more broadly.

Often, we think about human interaction with AI, but it can be useful to think of a three-way interaction with human(s), AI and plain-old computing – that is hand coded algorithms or classic AI. Now look at each kind of AI that you are thinking of using and ask what is it good for?

What kind of things do I mean? One of the problems with traditional AI is that it was good with hard-nosed rules, but much more problematic with fuzzy things.  There are various techniques such as Bayesian methods and fuzzy logic, but they require you to formalize the fuzziness into probabilities or similar functions.  Amongst other things this limited various forms of natural language understanding and common-sense reasoning

Of course, large language models are really good at dealing with the nuances of language, but LLMs are less good when they try to be very precise, not least because they keep hallucinating!

So as you design for AI, ask what is it good for, how can I use it most appropriately?

As an example of the appropriate use of AI,  my wife uses an app from “The Doctor’s Kitchen” (https://www.thedoctorskitchen.com) to help keep track of the health value of food.

You take a photo of a plate of food before you eat it and the app creates a report on its nutritional value: how much fibre and protein it contains and its inflammation index.  Is it likely to be good for you or bad?

You could imagine doing this by writing a complicated prompt to an LLM or train a deep learning algorithm with lots of plates of food and hand-curated reports.  The app does not work like that.

What it does is to use image processing AI to analyse the plate and work out what food is on it.  Indeed, you can press an edit button to see what it thinks you’ve got on your plate, and, if it’s got it wrong, edit it.  One assumes that a log of these edits helps to further train the image processing AI.

So the AI has been used for the fuzzy part of the task, working out that there are crisps on the plate but a no cake. It even manages to recognise hummus and estimate how much.  It is amazingly good, but does sometimes get things wrong in terms of the volume or even what is there; however, when that happens you can easily see and correct it.

So this is using AI for the fuzzy bit.

This table of contents will then go into a standard algorithm that uses tables of nutritional values to look up how much protein is in, say, 10 grams of almonds, add this up for the plate and hence generate the final nutritional report.

AI and traditional computing together — combining the two using the best aspects of each.

Note that this is more explainable, you know, what’s going on.

It is also more flexible in terms of you can choose to enhance different components and change others.

There is also less vendor tie in.  This is not removed entirely as you need a new AI to be retrained.  However, it is easier to swap just the food recognition part than if the whole system were in a single AI.

This is good from a business point of view, but it also means you are using less large-scale AI with its environmental, financial and democratically damaging effects, when you could be using simpler computation.

Coming next …

Part 8 – summary and recap

This final post will recap what we’ve learnt about the runaway nature of the AI industry, how it undermines free markets, and how we can make a difference. The core question is not what can AI do, but what should AI do?

 

Update

Since the talk in January I read about A.T.L.A.S. (Adaptive Test-time Learning and Autonomous Specialization) an AI coding system built by a business student Johnathon Tigges wanting to challenge the assumption that “only the biggest players can build meaningful things” [Ti26] .  It is able to outcompete the big coding agents by being clever – rather than just throwing a problem into a big code-optimised LLMs and asking for a solution, it uses AI to generate lots of potential code fragments and tests them, using the best to further refine the AI model … all on a consume GPU.  A lovely example of smart use of AI!  For a more detailed description see Sebastian Buzdugan’s Medium story about it [Bu26].

References

[BD23] Alba Bisante, Alan Dix, Emanuele Panizzi, and Stefano Zeppieri (2023). To err is AI.In Proceedings of the 15th Biannual Conference of the Italian SIGCHI Chapter, pp.1–11. https://dsoi.org/10.1145/3605390.3605414

[Bu26] Sebastian Buzdugan (2026). Why a $500 GPU Can Beat Claude Sonnet on Coding Benchmarks. Medium. Mar 28, 2026. https://medium.com/@sebuzdugan/why-a-500-gpu-can-beat-claude-sonnet-on-coding-benchmarks-6c8169ffe4fe

[DS24]  DeepSeek-AI (2024).  DeepSeek-V3 Technical Report. arXiv preprint. 27 Dec 2024. https://arxiv.org/abs/2412.19437

[DS25]  DeepSeek-AI (2025).  DeepSeek-V3. GitHub Repository. Release v1.0.0. 27 Jun 2025. https://github.com/deepseek-ai/DeepSeek-V3

[Di22] Dickson, B. (2022). Can large language models be democratized? TechTalk,-May 16, 2022. https://bdtechtalks.com/2022/05/16/opt-175b-large-language-models/

[DB00] A. Dix, R. Beale and A. Wood (2000).  Architectures to make Simple Visualisations using Simple Systems.  Proceedings of Advanced Visual Interfaces – AVI2000, ACM Press, pp. 51-60.  https://www.alandix.com/academic/papers/avi2000/

[Dx24] Alan Dix (2024). Patient Interaction – for well-being, productivity and sustainability. FUSION 2024, Kuala Lumpur, Malaysia, 28 Sept. 2024. https://www.alandix.com/academic/talks/FUSION2024/

[Dx25]  Dix, A. (2025). Artificial Intelligence – Humans at the Heart of Algorithms, 2nd Edition, Chapman and Hall.  https://alandix.com/aibook/

[bibitelm name=Dx26b] A. Dix. (2026). AI for Human–Computer Interaction. CRC Press, in press. https://alandix.com/ai4hci/

[Hi20] Hinton, G. (2020). Extrapolating the spectacular performance of GPT3 into the future suggests that the answer to life, the universe and everything is just 4.398 trillion parameters. Twitter (now X), Jun 10, 2020. https://x.com/geoffreyhinton/status/1270814602931187715

[HS22]  Hu, E. J., Shen, Y., et al. (2022). LoRa: Low-rank adaptation of large language models. ICLR, 1(2), 3. https://arxiv.org/abs/2106.09685

[LF24] Liu, A., Feng, B., et al. (2024). Deepseek-v3 technical report. arXiv preprint arXiv:2412.19437. https://arxiv.org/abs/2412.19437

[Sa23] Sajid, H. (2023).  Artificial Intelligence: Can You Build Large Language Models Like ChatGPT At Half Cost? Unite.ai, May 11, 2023.  https://www.unite.ai/can-you-build-large-language-models-like-chatgpt-at-half-cost/

[Ti26]  Johnathon Tigges (2026).  A.T.L.A.S. – Adaptive Test-time Learning and Autonomous Specialization. GitHub. https://github.com/itigges22/ATLAS

[TH23] Touvron, H., Martin, L., et al. (2023). Llama 2: Open foundation and fine-tuned chat models. arXiv preprint arXiv:2307.09288. https://arxiv.org/abs/2307.09288

[WP23] Wang, P., Panda, R., et al. (2023). Learning to grow pretrained models for efficient transformer training. arXiv preprint.  https://arxiv.org/abs/2303.00980

[We25b] Werner, J. (2025). Did DeepSeek Copy Off Of OpenAI? And What Is Distillation? Forbes, Jan 30, 2025. https://www.forbes.com/sites/johnwerner/2025/01/30/did-deepseek-copy-off-of-openai-and-what-is-distillation/

[ZD22]  Zhang, S., Diab, M. and Zettlemoyer, L. (2022). Democratizing access to large-scale language models with OPT-175B. Meta Research Blog, May 3, 2022. https://ai.meta.com/blog/democratizing-access-to-large-scale-language-models-with-opt-175b/

 

 

 

The Abomination of AI – part 6 – should we worry?

Posted on by

Runaway growth of AI is not painless – opportunity costs of investment and human costs of lost jobs.  Gains may be transitory – buy-now-pay-later tech risk tying users into spiralling costs.

This is the sixth of a series of blogs based on my keynote “The abomination of AI” at ICoSCI 2026.  Each has an accompanying segment of the video and slides from the talk as well as detailed notes and references.  Section numbers refer to the full report which will be released in the final blog.   The slide thumbnails in the text correspond to the slides in the navigation panel below.  The presentation can be played below, or opened full screen. The full length video, complete slides and further information can be found at: https://alandix.com/academic/talks/ICOSCI-2026-abomination-of-AI/

Previously …

§1.  Every industry is driven by profits and power, but there is something about the nature of AI itself, which interacts with the nature of market forces in the world that is problematic and is different from other technologies.

§2.  Can any technology be neutral?  AI can be used for good purposes, such as advances in healthcare.  It can also have bad outcomes such as bias in the criminal justice system or online exploitative pornography.  Perhaps most often it is creating the frivolous or even ugly.

§3.  The obvious impact of AI is in the things it does directly. Some technologies also change the very nature of society, affecting even those who do not use them. Cars are an obvious example.  AI is also such a technology.

§4.  Doomsayers worry about the point when AI becomes sentient, outgrowing its creators.  The real danger is more insidious: the massive financial and human impacts of AI seem almost obscene.

§5 .   Network externalities, the way one person’s use of AI and digital tech changes its value for others, creates positive feedback loops, leading to runaway growth and emergent monopolies, the nemesis of free markets. This the very nature of digital technology and AI breaks free markets leading to runaway inequality, even with the best intentions of industry … but some tech companies further exploit these effects.

6.  Should we worry?

6.1  Jobs and power

Image: Scottish Government, CC BY 2.0. https://commons.wikimedia.org/wiki/File:One_of_the_typing_pools_%283829002585%29.jpg

Does this matter?  So what if a small number of companies have notional multi-trillion balance sheets and are engaged in runaway development in the digital realm, so long as it doesn’t affect the real world.  But, of course, it does; the digital domain is leaking into the physical domain.

Of course, technology and automation have long had massive impact on society, with a gradual shift from human expertise to financial capital.  This certainly dates back to the 19th century or late 18th century with the rise of the industrial revolution.  Of course, at that time, humans were still needed, but they went from being the expert weavers and spinners, to the those (including young children) who merely tended the machines, monitoring and knotting broken threads, and occasionally losing arms in the moving parts.  So it wasn’t that the humans were unnecessary, but they merely fed the machines.

Moving into the 20th century, machines replaced humans more completely with fully automated production lines and industrial robots, although of course still with humans cleaning up between them.  In many parts of the global north skilled manual work has all but disappeared, with a combination of automation and out-sourcing.

To some extent the impact of automation initially hit traditional male jobs, but in the latter half of 20th century, from about the early seventies on this also hit clerical roles.  Until then every big organisation would have had a typing pool.  My own mother was for many years a typist at first in the War Department throughout the Second World War, and then the Inland Revenue. These typing pools, consisted of ranks of people, usually women, typing sometimes from dictation and shorthand, and sometimes other forms of handwriting.  Word processors basically destroyed the typing pool. Whereas previously, managers would dictate letters and reports to a secretary who would then type it up,  with the word processor, despite initial resistance, they would type directly themselves.  Of course, after initial diversity, Microsoft Word soon became dominant – another emergent monopoly, although now matched by Google Workspace, with about 90% world share between them.

So, in general, skilled working class jobs have been destroyed by automation leaving a growing underclass with minimum wage jobs and gig work.

What we’re seeing now is that the mid-range intellectual work is starting to be eaten by AI [BSI25,].

You may have seen the MIT report that found that while many companies were investing heavily in AI, around 95% of the projects were considered to be failing or underperforming [CP25].  So, effective job substitution is not yet universal, but in some areas such as computing many of the lower range of the roles, typical graduate first jobs, are being replaced by AI.  Until recently the expert developer would have several junior developers who do the grunt work; now this is done by AI.  Similar pictures are emerging in advertising, aspects of finance, and some of the large management consultancies [Ko25,Sw26,IPA26,KM26,Pa26].  In the UK, and even more so in other parts of the world, there are strong pushes to use AI more extensively within government, not least on the assumption that it will improve efficiency [GUK25,Dx26].

There’s a critical issue about who’s in control.  Think about the road network.  In the UK there are some private roads and also some toll roads, but the majority of roads, including almost all in urban areas, are owned by the local authority or central government.  That is, the vast majority of the road network is local in terms of its maintenance and control.  Imagine if the road network was instead owned by two or three major companies based in the west coast of America.  Imagine if every road in Malaysia, every road in Indonesia, as well as every road in UK was owned by those two or three companies  half the world away.  If there’s a pothole in the road, it is those companies to whom you have to complain.  Perhaps they decide to charge you to use the road outside your house or decide to remove the roads entirely if they’re in dispute with you or your government.

That’s exactly the direction we are moving with AI and public services.  Even assuming the best intention of the big AI players, this does feel worrying.  And, of course, this isn’t a choice you can make or not.  Just like cars and roads, once AI is embedded into public service everything orients around it.

Returning to the changes in employment, once we lose the entry stage jobs, there’s a clear problem for the people who would have had them.  All the graduates from our universities, who would have been going into those jobs, are being hit, and, in many countries, on top of large student loans [DoE25,Pa25,Pa26].  This is creating a class of people who are underemployed, inexperienced, and quite likely disaffected with society.  Think of this in the light of the rise of extremism across the world.  Often this is dismissed as a problem of the uneducated, but here we are adding a vast number of highly educated people, who are disaffected in society, further spreading those extreme messages.

 

6.2  Locked into AI

This is also a problem within an organization.  If you are not employing those early career people, what happens in five or ten years’ time as your more experienced employees want to move up the organization?  How do you fill in those gaps if you haven’t been training people?

This might be something we need to address as universities, training people effectively to higher and higher levels so that they can jump in at that point.

Or the organisation can simply find they need more AI – what they certainly can’t do is just turn off the AI because they haven’t got the people with the experience in order to do the jobs anymore.  They have become locked in as a company to the use of AI.

This is also true of data.  Microsoft have a guide entitled, “Prepare your data for AI” [Ms26].  The use of AI is not coming for free, but needs a rearrangement of data for it.  One does wonder if the same effort making data ready for AI could be better spent making it ready for simpler statistical algorithms.

However, let’s assume you have put effort into reorienting your whole data around AI. Your systems rapidly become AI dependent – your recent information and new data has become deeply embedded into the AI itself in ways that are often opaque.

Once you have bought into an AI system, you can’t just say, “well, let’s just swap to something else”.  It’s difficult even to swap vendors once it is that embedded.

 

6.3  Buy now … pay later

If you have a loan with interest, you know you have to pay for it eventually, but things can be less obvious.  When I was little, my mum had a Kays catalogue, a sort of the 1960s  equivalent of internet selling [WA17].  Its pages were full of big colour pictures of clothes, white goods, toys, etc. …it was usually the toys I was looking at.  You could buy things from the catalogue and could pay over 20 weeks with no interest, but of course the things cost more than if you had the ready cash buy them at a shop.  So effectively you were paying extra.

AI currently is in that ‘buy now pay later’ mode, both globally and locally for individuals.  AI growth is funded by massive investment (as we discussed absolutely huge) possibly more than ever before except perhaps the South Sea Bubble.  However, the income doesn’t in any way cover the costs, and the ratio between expected income and investment is way out of kilter for what you’d expect even for a digital company, let alone for a physical one.

So how do the books add up?

If you’re an accountant in the company or if you’re an investment manager, what are you thinking about as, as you see these figures?  Why don’t you sound the alarm?  The reason is you are thinking that in the future you will have more money from that stream.  In early digital companies, like Amazon, you did that because you assumed you were going have a bigger market, the number of people who would use it would grow.

But AI already has lots of users, so instead you have two options.  The first is to find ways to make what you produce more cheaply, which is happening to some extent already. However, you don’t want it to get too cheap otherwise competitors can enter the market.  The alternative, and your only real option, to recoup your investment by charging more or getting the same customers to use more. Either way, it is the customer who pays in the end!

This is no secret.  Fortune magazine said that OpenAI’s business plan relies on “what amounts to a bet on dominance” [Sm25].  That is, in putting in all that investment, what investors are hoping is that the company will become the AI company in an area that everybody is tied into.  And then of course they can charge pretty much what they like: a buy now – pay later world. We’re using AI now, but the cost is going to come later on.

 

Coming next …

Part 7 – what can we do?

It all seems too big, requiring national and international responses.  But we can make a difference using appropriately chosen small AI (including none). Plus, this good use of AI is good for business too.

 

Update.

Since the talk, I read about a woman who had developed a close relationship with a chatbot hosted on a version of ChatGPT that is due to be retired [He26]. While she could probably export her chat history and use that to reinitialise the new version of the software, it would not be the same.  We will soon start to hear similar stories for business and public systems as tech companies have not had a good record of backward compatibility, and this is all but impossible with current LLMs.

Also, in late January, OpenClaw was released [OC26].  This highlighted the way current payment models do not reflect the actual cost of use. OpenClaw (originally called Clawdbot) is an open-source GitHub project that used the Claude API to create an automated assistant coordinating web and desktop resources.  Within days of the launch Anthropic enforced a long-standing, but unenforced, restriction on third-party use of its API and blocked OpenClaw for most user accounts including its $200 Max account.  This was because these accounts come with monthly usage limits, and OpenClawd encouraged full use of those limits.  However, the business model of even premium accounts depends on users NOT using their monthly allowances. OpenClawd encouraged full use of those limits, thus exposing.the true cost of the full use vastly exceeded the subscriptions [Ba26] .

 

 

References

[Ba26] Novy Baf (2026).  Anthropic Pushed Its Most Loyal Developers Straight Into OpenAI’s arms. OpenAI Didn’t Even Have to Ask.  The Nov TEch, 2nd Mar 2026.  https://www.thenovtech.com/p/anthropic-pushed-its-most-loyal-developers

[BSI25] British Standards Institution (2025). Evolving Together: AI, automation and  building the skilled  workforce of the future.  https://www.bsigroup.com/en-GB/insights-and-media/insights/whitepapers/evolving-together-flourishing-in-the-ai-workforce/

[Dx26] A. Dix. (2026). Beyond the Algorithm: Designing Human-Centric Public Service with AI. Talk at Service Design for Public Sector Spotlight Seminar series of challenges and opportunities between Design Cultures and Public Sector, Sapienza, University of Rome + Online, 4th February 2026. https://alandix.com/academic/talks/Rome-Seminar-Feb-2026/

[DoE25] Department of Education (2025). The impact of AI on UK jobs and training. November 2023.  https://www.gov.uk/government/publications/the-impact-of-ai-on-uk-jobs-and-training

[CP25]  Aditya Challapally, Chris Pease, Ramesh Raskar, Pradyumna Chari (2025). The GenAI Divide: State of AI in Business 2025. MIT NANDA, July 2025. https://mlq.ai/media/quarterly_decks/v0.1_State_of_AI_in_Business_2025_Report.pdf

[GUK25] Gov.UK (2025). AI to power national renewal as government announces billions of additional investment and new plans to boost UK businesses, jobs and innovation. Press release from Department for Science, Innovation and Technology, HM Treasury, Wales Office, The Rt Hon Liz Kendall MP, The Rt Hon Rachel Reeves MP and The Rt Hon Jo Stevens MP.  20 November 2025. https://www.gov.uk/government/news/ai-to-power-national-renewal-as-government-announces-billions-of-additional-investment-and-new-plans-to-boost-uk-businesses-jobs-and-innovation

[He26]  Stephanie Hegarty (2026). Rae fell for a chatbot called Barry, but their love might die when ChatGPT-4o is switched off. BBC News, 14 February 2026. https://www.bbc.co.uk/news/articles/crl43dxwwy9o

[IPA26] IPA (2026). IPA Agency Census 2025 shows workforce declines while diversity improves.  Institute of Practitioners in Advertising. 11 February 2026. https://ipa.co.uk/news/agency-census-2025/

[KM26] Lucy Knight and Sumaiya Motara (2026). The big AI job swap: why white-collar workers are ditching their careers. The Guardian,  11 Feb 2026. https://www.theguardian.com/technology/2026/feb/11/big-ai-job-swap-white-collar-workers-ditching-their-careers

[Ko25] Saskia Koopman (2025).  Big Four slash graduate jobs as AI takes on entry level work. City AM, 23 June 2025. https://www.cityam.com/big-four-slash-graduate-jobs-as-ai-takes-on-entry-level-work/

[Ms26] Microsoft (2026). Prepare your data for AI. Dated 20/1/2026.  https://learn.microsoft.com/en-gb/power-bi/create-reports/copilot-prepare-data-ai

[OC26] OpenClaw (2026).  OpenClaw — Personal AI Assistant. https://github.com/openclaw/openclaw

[Pa25] Joanna Partridge (2025). Gen Z faces ‘job-pocalypse’ as global firms prioritise AI over new hires, report says. The Guardian,  9 Oct 2025. https://www.theguardian.com/money/2025/oct/09/gen-z-face-job-pocalypse-as-global-firms-prioritise-ai-over-new-hires-report-says

[Pa26] Joanna Partridge (2026). More than a quarter of Britons say they fear losing jobs to AI in next five years. The Guardian,  25 Jan 2026. https://www.theguardian.com/business/2026/jan/25/more-than-quarter-britons-fear-losing-jobs-ai-next-five-years

[Sm25]  Dave Smith (2025). OpenAI says it plans to report stunning annual losses through 2028—and then turn wildly profitable just two years later . Fortune, November 12, 2025. https://fortune.com/2025/11/12/openai-cash-burn-rate-annual-losses-2028-profitable-2030-financial-documents/

[Sw26] Mark Sweney (2026). UK ad agencies undergo their biggest exodus of staff as AI threatens industry. The Guardian,  13 Feb 2026. https://www.theguardian.com/media/2026/feb/13/uk-ad-agencies-biggest-annual-exodus-of-staff-ai-threatens-industry

[WA17]  Worcestershire Archive and Archaeology Service (2017).  Christmas and Kays.  Explore the Past. 19th December 2017. https://www.explorethepast.co.uk/2017/12/christmas-and-kays/

 

 

 

 

Retiring today!

Posted on by

Today is my last day at Cardiff Met.  I retired from Swansea University just over a year ago, and continued my part-time position at Cardiff Met at that point, but now I am completing my retirement.  I will continue to be emeritus professor at both Swansea and Cardiff Met, so I’m not turning my back on academia; indeed I’ve got a second edition of “Introduction to Artificial Intelligence” due out in June and I’m part way through several other books, so I won’t be idle.  However, it is a change of pace and focus.

To mark my retirement and also turning 65 in July I am running 65 miles in a week in aid of Cancer Research UK.  Read more about this at 65 for 65.

It has a been a joy working with everyone at Cardiff Met for the past three years.  I’ve been part of the central Research and Innovation Services where I’ve sometimes described myself as a ‘professor without portfolio’ and that my job was to ‘talk to anyone, anywhere, about anything’.  Although slightly tongue in cheek, the latter has been largely true, spending time with researchers and academics in areas from creative writing to bacterial infection.  So many wonderful people each with a vision to make a difference in the world.

My academic connections with Cardiff Met go back to the mid 2000s when I started to collaborate with Steve Gill and others in product design and I’d been a visiting professor in the Cardiff School of Art and Design since 2013.  However, the roots reach much further back.

TouchIT front cover

TouchIT: Understanding Design
in a Physical-Digital World
A. Dix, S. Gill, D. Ramduny-Ellis & J Hare

Public Lecture.
Treading-Out Technology
Visiting Professor Alan Dix

When I was in Sixth Form at Howardian Comprehensive School (now closed and demolished), once a week, every Friday afternoon, a friend and I used to walk several miles across Cardiff to the FE college on Western Avenue to study for a computing A Level … the FE College that many years later became the Llandaff Campus of Cardiff Metropolitan University.

Llandaff Technical College opened in December 1954.

According to a BBC article the first computing A Level exams were in 1971, but even in 1976–78 it was still a new subject and not one the school could teach.  I don’t recall the name of the lecturer, who I’m guessing was fresh out of university and found the A Level gave him some creative room compared with the City and Guilds courses that were the staple of the college.  This was in the days before the crazy gulf we see now between more theoretical and practical aspects of the subject.  Each week was something different, from the innards of machines to application areas, the halting problem and even an afternoon of COBOL programming.  Indeed we covered more programming languages in those couple of hours a week than are common in degree courses today.  My own coursework project was writing a bootstrap compiler.

Crucially the A Level gave me access to the FE College library and in particular maths books, many of the “X for engineers” style, which are usually so much better written than the equivalent ones aimed at mathematicians!  This was invaluable when I later went on to sit the Cambridge maths entrance exam, won a scholarship and was then contacted by the British Maths Olympiad organisers, which led to being part of the UK Team at the International Mathematical Olympiad in Bucharest, my first time out of the country.

And it all started in the buildings that I have been working in for the past few years.

Query-by-Browsing gets local explanations

Posted on by

Query-by-Browsing (QbB) now includes local explanations so that you can explore in detail how the AI generated query relates to dataset items.

Query-by-Browsing is the system envisaged in my 1992 paper that first explored the dangers of social, ethnic and gender bias in machine-learning algorithms.  QbB generates queries, in SQL or decision tree form based on examples of records that the user does or does not want. A core feature has always been the dual intensional (query) and extensional (selected data) to aid transparency.

QbB has gone through various iterations and a simple web version has been available for twenty years, and was updated last year to allow you to use your own data (uploaded as CSV files) as well as the demo datasets.

The latest iteration also includes a form of local explanation.  If you hover over a row in the data table it shows which branch of the query meant that the row was either selected or not.

Similarly hovering over the query shows you which data rows were selected by the query branch.

However, this is not the end of the story!

In about two weeks Tommaso will be presenting our paper “Talking Back: human input and explanations to interactive AI systems” at the Workshop on Adaptive eXplainable AI (AXAI) at IUI 2025 in Cagliari, Italy,  A new version of QbB will be released to coincide with this.  This will include ‘user explanations’, allowing the user to tell the system why certain records are important to help the machine learning make better decisions.

Watch this space …

 

 

 

Another year – running and walking, changing roles and new books

Posted on by

Yesterday I completed the Tiree Ultramarathon, I think my sixth since they began in 2014. As always a wonderful day and a little easier than last year. This is always a high spot in the year for me, and also corresponds to the academic year change, so a good point to reflect on the year past and year ahead.  Lots of things including changing job role, books published and in preparation, conferences coming to Wales … and another short walk …

Tiree Ultra and Tech Wave

Next week there will be a Tiree Tech Wave, the first since Covid struck. Really exciting to be doing this again, with a big group coming from Edinburgh University, who are particularly interested in co-design with communities.

Aside: I nearly wrote “the first post-Covid Tiree Tech Wave”, but I am very aware that for many the impact of Covid is not past: those with long Covid, immunocompromised people who are in almost as much risk now as at the peak of the pandemic, and patients in hospital where Covid adds considerably to mortality.

Albrecht Schmidt from Ludwig-Maximilians-Universität München was here again for the Ultra. He’s been several times after first coming the year of 40 mile an hour winds and rain all day … he is built of stern stuff.  Happily, yesterday was a little more mixed, wind and driving rain in the morning and glorious sunshine from noon onwards … a typical Tiree day 😊

We have hatched a plan to have Tiree Tech Wave next year immediately after the Ultra. There are a number of people in the CHI research community interested in technology related to outdoors, exercise and well-being, so hoping to have that as a theme and perhaps attract a few of the CHI folk to the Ultra too.

Changing roles

My job roles have changed over the summer.

I’ve further reduced my hours as Director of the Computational Foundry to 50%. University reorganisation at Swansea over the last couple of years has created a School of Mathematics and Computer Science, which means that some of my activities helping to foster research collaboration between CS and Maths falls more within the School role. So, this seemed a good point to scale back and focus more on cross-University digital themes.

However, I will not be idle! I’ve also started a new PT role as Professorial Fellow at Cardiff Metropolitan University. I have been a visiting professor at the Cardiff School of Art and Design for nearly 10 years, so this is partly building on many of the existing contacts I have there. However, my new role is cross-university, seeking to encourage and grow research across all subject areas. I’ve always struggled to fit within traditional disciplinary boundaries, so very much looking forward to this.

Books and writing

This summer has also seen the publication of “TouchIT: Understanding Design in a Physical-Digital World“. Steve, Devina, Jo and I first conceived this when we were working together on the DePTH project, which ran from 2007 to 2009 as part of the AHRC/EPSRC funded Designing for the 21st Century Initiative. The first parts were written in 2008 and 2009 during my sabbatical year when I first moved to Tiree and Steve was our first visitor. But then busyness of life took over until another spurt in 2017 and then much finishing off and updating. However now it is at long last in print!

Hopefully not so long in the process, three more books are due to be published in this coming year, all around an AI theme. The first is a second edition of the “Introduction to Artificial Intelligence” textbook that Janet Finlay and I wrote way back in 1996. This has stayed in print and even been translated into Japanese. For many years the fundamentals of AI only changed slowly – the long ‘AI winter’. However, over recent years things have changed rapidly, not least driven by massive increases in computational capacity and availability of data; so it seemed like a suitable time to revisit this. Janet’s world is now all about dogs, so I’ve taken up the baton. Writing the new chapters has been easy. The editing making this flow as a single volume has been far more challenging, but after a focused writing week in August, it feels as though I’ve broken the back of it.

In addition, there are two smaller volumes in preparation as part of the Routledge and CRC AI for Everything series. One is with Clara Crivellaro on “AI for Social Justice“, the other a sole-authored “AI for Human–Computer Interaction”.

All of these were promised in 2020 early in the first Covid lockdown, when I was (rather guiltily) finding the time tremendously productive. However, when the patterns of meetings started to return to normal (albeit via Zoom), things slowed down somewhat … but now I think (hope!) all on track 😊

Welcoming you to Wales

In 2023 I’m chairing and co-chairing two conferences in Swansea. In June, ACM Engineering Interactive Computer Systems (EICS 2023) and in September the European Conference on Cognitive Ergonomics (web site to come, but here is ECCE 2022). We also plan to have a Techwave Cymru in March. So I’m looking forward to seeing lots of people in Wales.

As part of the preparation to EICS I’m planning to do a series of regular blog posts on more technical aspects of user interface development … watch this space …

Alan’s on the road again

Nearly ten years ago, in 2013, I walked around Wales, a personal journey and research expedition. I always assumed I would do ‘something else’, but time and life took over. Now, the tenth anniversary is upon me and it feels time do something to mark it.

I’ve always meant to edit the day-by-day blogs into a book, but that certainly won’t happen next year. I will do some work on the dataset of biodata, GPS, text and images that has been used in a few projects and is still a unique data set, including, I believe, still the largest single ECG trace in the public domain.

However, I will do ‘something else’.

When walking around the land and ocean boundaries of Wales, I was always aware that while in some sense this ‘encompassed’ the country, it was also the edge, the outside. To be a walker is to be a voyeur, catching glimpses, but never part of what you see.  I started then to think of a different journey, to the heart of Wales, which for me, being born and brought up in Cardiff, is the coal valleys stretching northwards and outwards. The images of coal blackened miners faces and the white crosses on the green hillside after Aberfan are etched into my own conception of Wales.

So, there will be an expedition, or rather as series of expeditions, walking up and down the valleys, meeting communities, businesses, schools and individuals.

Do you know places or people I should meet?

Do you want to join me to show me places you know or to explore new places?

Busy September – talks, tutorials and an ultra-marathon

Posted on by

September has been a full month!

During the last two weeks things have started to kick back into action, with the normal rounds of meetings and induction week for new students.  For the latter I’d pre-recorded a video welcome, so my involvement during the week was negligible.  However, in addition I delivered a “Statistics for HCI” day course organised by the BCS Interaction Group with PhD students from across the globe and also a talk “Designing User Interactions with AI: Servant, Master or Symbiosis” at the AI Summit London.  I was also very pleased to be part of the “60 faces of IFIP” campaign by the International Federation for Information Processing.

It was the first two weeks that stood out though, as I was back on Tiree for two whole weeks.  Not 100% holiday as during the stay I gave two virtual keynotes: “Qualitative–Quantitative Reasoning: thinking informally about formal things” at the International Colloquium on Theoretical Aspects of Computing (ICTAC) in Kazakhstan and “Acting out of the Box” at the University of Wales Trinity St David (UWTSD) Postgraduate Summer School.  I also gave a couple of lectures on “Modelling interactions: digital and physical” at the ICTAC School which ran just before the conference and presented a paper on “Interface Engineering for UX Professionals” in the Workshop on HCI Engineering Education (HCI-E2) at INTERACT 2021 in Bari.  Amazing how easy it is to tour the world from a little glamping pod on a remote Scottish Island.

Of course the high point was not the talks and meetings, but the annual Tiree Ultra-marathon.  I’d missed last year, so especially wonderful to be back: thirty five miles of coastline, fourteen beaches, not to mention so many friendly faces, old friends and new.  Odd of course with Covid zero-contact and social distancing – the usual excited press of bodies at the pre-race briefing in An Talla, the Tiree community hall, replaced with a video webinar and all a little more widely spaced for the start on the beach too.

The course was slightly different too, anti-clockwise and starting half way along Gott Bay, the longest beach.  Gott Bay is usually towards the end of the race, about 28 miles in, so the long run, often into the wind is one of the challenges of the race.  I recall in 2017 running the beach with 40 mile an hour head wind and stinging rain – I knew I’d be faster walking, but was determined to run every yard of beach..  Another runner came up behind me and walked in my shelter.  However, this year had its own sting in the tail with Ben Hynish, the highest point, at 26 miles in.

The first person was across the line in about four-and-a-quarter hours, the fastest time yet.  I was about five hours later!

This was my fifth time doing the ultra, but the hardest yet, maybe in part due to lockdown couch-potato-ness!  My normal training pattern is that about a month before the ultra I think, “yikes, I’ve not run for a year” and then rapidly build up the miles – not the recommended training regime!  This year I knew I wasn’t as fit as usual, so I did start in May … but then got a knee injury, then had to self-isolate … and then it was into the second-half of July; so about a month again.

Next year it will be different, I will keep running through the winter … hmm … well, time will tell!

The different September things all sound very disparate – and they are, but there are some threads and connections.

The first thread is largely motivational.

The UWTSD keynote was about the way we are not defined by the “kind of people” we think of ourselves as being, but by the things we do.  The talk used my walk around Wales in 2013 as the central example, but the ultra would have been just as pertinent.  Someone with my waistline is not who one would naturally think as being an ultramarathon runner – not that kind of person, but I did it.

However, I was not alone.  The ‘winners’ of the ultra are typically the rangy build one would expect of a long-distance runner, but beyond the front runners, there is something about the long distance that attracts a vast range of people of all ages, and all body shapes imaginable.  For many there are physical or mental health stories: relationship breakdowns, illnesses, that led them to running and through it they have found ways to believe in themselves again.  Post Covid this was even more marked: Will, who organises the ultra, said that many people burst into tears as they crossed the finish line, something he’d never seen before.

The other thread is about the mental tools we need to be a 21st century citizen.

The ICTAC keynote was about “Qualitative–Quantitative Reasoning”, which is my term for the largely informal understanding of numbers that is so important for both day-to-day and professional life, but is not part of formal education.  The big issues of our lives from Covid to Brexit to climate change need us to make sense of large-scale numerical or data-rich phenomena.  These often seem too complex to make sense of, yet are ones where we need to make appropriate choices in both our individual lives and political voices.  It is essential that we find ways to aid understanding in the public, press and politicians – including both educational resources and support tools.

The statistics course and my “Statistics for HCI” book are about precisely this issue – offering ways to make sense of often complex results of statistical analysis and obtain some of the ‘gut’ understanding that professional statisticians develop over many years.

My 60 faces of IFIP statement also follows this broad thread:

“Digital techology is now essential to being a citizen. The future of information processing is the future of everyone; so needs to be understood and shaped by all. Often ICT simply reinforces existing patterns, but technology is only useful if we can use it to radically reimagine a better world.”

More information on different events

Tiree Ultra

Tiree Ultramarathon web page and Facebook Group

Paper: Interface Engineering for UX Professionals

HCI-E2: Workshop on HCI Engineering Education – for developers, designers and more, INTERACT 2021, Bari, Italy – August 31st, 2021. See more – paper and links

Summer School Lecturea: Modelling interactions: digital and physical

Lecture at ICTAC School 2021: 18th International Colloquium on Theoretical Aspects of Computing, Nazarbayev University, Nur-Sultan, Kazakhstan, 1st September 2021. See more – abstract and links

Talk: Designing User Interactions with AI: Servant, Master or Symbiosis

The AI Summit London, 22nd Sept. 2021. See more – abstract and links

Day Course: Statistics for HCI

BCS Interaction Group One Day Course for PhD Students, 21st Sept. 2021.
See my Statistics for HCI Micro-site.

Keynote: Acting out of the Box

Rhaglen Ysgol Haf 2021 PCYDDS / UWTSD Postgraduate Summer School 2021, 10th Sept. 2021. See more – abstract and links

Keynote: Qualitative–Quantitative Reasoning: thinking informally about formal things

18th International Colloquium on Theoretical Aspects of Computing, Nazarbayev University, Nur-Sultan, Kazakhstan, 10th Sept. 2021. See more – full paper and links

Induction week greeting

 

Darwinian markets and sub-optimal AI

Posted on by

Do free markets generate the best AI?  Not always, and this not only hits the bottom line, but comes with costs for personal privacy and the environment.  The combination of network effects and winner-takes-all advertising expenditure means that the resulting algorithms may be worst for everyone.

A few weeks ago I was talking with Maria Ferrario (Queens University Belfast) and Emily Winter (Lancaster University) regarding privacy and personal data.  Social media sites and other platforms are using ever more sophisticated algorithms to micro-target advertising.  However, Maria had recently read a paper suggesting that this had gone beyond the point of diminishing returns: far simpler  – and less personally intrusive – algorithms achieve nearly as good performance as the most complex AI.  As well as having an impact on privacy, this will also be contributing to the ever growing carbon impact of digital technology.

At first this seemed counter-intuitive.  While privacy and the environment may not be central concerns, surely companies will not invest more resources in algorithms than is necessary to maximise profit?

However, I then remembered the peacock tail.


Jatin Sindhu, CC BY-SA 4.0, via Wikimedia Commons
The peacock tail is a stunning wonder of nature.  However, in strict survival terms, it appears to be both flagrantly wasteful and positively dangerous – like eye-catching supermarket packaging for the passing predator.

The simple story of Darwinian selection suggests that this should never happen.  The peacocks that have smaller and less noticeable tails should have a better chance of survival, passing their genes to the next generation, leading over time to more manageable and less bright plumage.  In computational terms, evolution acts as a slow, but effective optimisation algorithm, moving a population ever closer to a perfect fit with its environment.

However, this simple story has twists, notably runaway sexual selection.  The story goes like this.  Many male birds develop brighter plumage during mating season so that they are more noticeable to females.  This carries a risk of being caught by a predator, but there is a balance between the risks of being eaten and the benefits of copulation.  Stronger, faster males are better able to fight off or escape a predator, and hence can afford to have slightly more gaudy plumage.  Thus, for the canny female, brighter plumage is a proxy indicator of a more fit potential mate.  As this becomes more firmly embedded into the female selection process, there is an arms race between males – those with less bright plumage will lose out to those with brighter plumage and hence die out.  The baseline constantly increases.

Similar things can happen in free markets, which are often likened to Darwinian competition.

Large platforms such as Facebook or Google make the majority of their income through advertising.  Companies with large advertising accounts are constantly seeking the best return on their marketing budgets and will place ads on the platform that offers the greatest impact (often measured by click-through) for the least expenditure.  Rather like mating, this is a winner-takes-all choice.  If Facebook’s advertising is 1% more effective than Google’s  a canny advertiser will place all their adverts with Facebook and vice versa.  Just like the peacock there is an existential need to outdo each other and thus almost no limit on the resources that should be squandered to gain that elusive edge.

In practice there are modifying factors; the differing demographics of platforms mean that one or other may be better for particular companies and also, perhaps most critically, the platform can adjust its pricing to reflect the effectiveness so that click-through-per-dollar is similar.

The latter is the way the hidden hand of the free market is supposed to operate to deliver ‘optimal’ productivity.  If spending 10% more on a process can improve productivity by 11% you will make the investment.  However, the theory of free markets (to the extent that it ever works) relies on an ‘ideal’ situation with perfect knowledge, free competition and low barriers to entry.  Many countries operate collusion and monopoly laws in pursuit of this ‘ideal’ market.

Digital technology does not work like this. 

For many application areas, network effects mean that emergent monopolies are almost inevitable.  This was first noticed for software such as Microsoft Office – if all my collaborators use Office then it is easier to share documents with them if I use Office also.  However, it becomes even more extreme with social networks – although there are different niches, it is hard to have multiple Facebooks, or at least to create a new one – the value of the platform is because all one’s friends use it.

For the large platforms this means that a substantial part of their expenditure is based on maintaining and growing this service (social network, search engine, etc.).  While the income is obtained from advertising, only a small proportion of the costs are developing and running the algorithms that micro-target adverts.

Let’s assume that the ratio of platform to advertising algorithm costs is 10:1 (I suspect it is a lot greater).  Now imagine platform P develops an algorithm that uses 50% more computational power, but improves advertising targeting effectiveness by 10%; at first this sounds a poor balance, but remember that 10:1 ratio.

The platform can charge 10% more whilst being competitive.   However, the 50% increase in advertising algorithm costs is just 5% of the overall company running costs, as 90% are effectively fixed costs of maintaining the platform.  A 5% increase in costs has led to a 10% increase in corporate income.  Indeed one could afford to double the computational costs for that 10% increase in performance and still maintain profitability.

Of course, the competing platforms will also work hard to develop ever more sophisticated (and typically privacy reducing and carbon producing) algorithms, so that those gains will be rapidly eroded, leading to the next step.

In the end there are diminishing returns for effective advertising: there are only so many eye-hours and dollars in users’ pockets. The 10% increase in advertising effectiveness is not a real productivity gain, but is about gaining a temporary increase in users’ attention, given the current state of competing platforms’ advertising effectiveness.

Looking at the system as a whole, more and more energy and expenditure are spent on algorithms that are ever more invasive of personal autonomy, and in the end yield no return for anyone.

And it’s not even a beautiful extravagance.

Free AI book and a new one coming …

Posted on by

Yes a new AI book is coming … but until then you can download the first edition for FREE 🙂

Many years ago Janet Finlay and I wrote a small introduction to artificial intelligence.  At the time there were several Bible-sized tomes … some of which are still the standard textbooks today.  However, Janet was teaching a masters conversion course and found that none of these books were suitable for taking the first steps on an AI journey, especially for those coming from non-computing disciplines.

Over the years it faded to the back of our memories, with the brief exception of the time when, after we’d nearly forgotten it, CRC Press issued a Japanese translation.  Once or twice the thought of doing an update arose, but quickly passed.  This was partly because our main foci were elsewhere, but also, at the danger of insulting all my core-AI friends, not much changed in core AI for many years!

Coming soon … Second Edition

Of course over recent years things have changed dramatically, hence my decision, nearly 25 years on, to create a new edition maintaining the aim to give a rich but accessible introduction, but capturing some of the recent trends and giving these a practical and human edge.  Following the T-model of teaching, I’d like to help both newcomer and expert gain a broad perspective of the issues and landscape, whilst giving enough detail for those that want to delve into a more specific area.

A Free Book and New Resources

In the mean time the publisher, Taylor & Francis/CRC has agreed to make the PDF of the first edition available free of charge  I have updated some of the code examples from the first edition and will be incrementally adding new material to the second edition micro-site including slides, cases studies, video and interactive materials.  If you’d like to teach using this please let me know your views on the topics and also if there are areas where you’d like me to create preliminary material with greater urgency.  I won’t promise to be able to satisfy everyone, but can use this to adjust my priorities.

Why now?

The first phase of change in AI was driven by the rise of big data and the increasing use of forms of machine learning to drive adverts, search results and social media.  Within user interface design, many of the fine details of colour choices and screen layout are now performed using A–B testing …sight variants of interfaces delivered to millions of people – shallow, without understanding and arguably little more than bean counting, but in numerous areas vast data volume has been found to be ‘unreasonably effective‘ at solving problems that were previously seen to be the remit of deep AI.

In the last few years deep learning has taken over as the driver of AI research and often also media hype.  Here it has been the sheer power of computation, partly due to Moores’ Law with computation nearly a million times faster than it was when that first edition was written nearly 25 years ago.  However, it has also been enabled by cloud computing allowing large numbers of computers ti efficiently attack a single problem.  Algorithms that might have been conceived of but dismissed as impractical in the past have become commonplace.

Alongside this has been a dark side of AI, from automated weapons and mass surveillance, to election rigging and the insidious knowledge that large corporations have gathered through our day-to-day web interactions.  In the early 1990s I warned of the potential danger of ethnic and gender bias in black-box machine learning and I’ve returned to this issue more recently as those early predictions have come to pass.

Across the world there are new courses running or being planned and people who want to know more.  In Swansea we have a PhD programme on people-first AI/big data, and there is currently a SIGCHIItaly workshop call out for Teaching HCI for AI: Co-design of a Syllabus. There are several substantial textbooks that offer copious technical detail, but can be inaccessible for the newcomer or those coming from other disciplines.  There are also a number of excellent books that deal with the social and human impact of AI, but without talking about how it works.

I hope to be able to build upon the foundations that Janet and I established all those years ago to create something that fills a crucial gap: giving a human-edge to those learning artificial intelligence from a computing background and offering an accessible technical introduction for those approaching the topic from other disciplines.