Digital technology seems to lead inexorably towards the concentration of power into a tiny number of vast corporations. As physical manufacture is increasingly digital and AI-based, will the same happen – vast factories of robots where humans remain only to sweep the floors? Is there any hope for the small company let alone individual craftsperson? The maker movement has led the way with open frameworks including open hardware construction, which extends into home automation for the enthusiast. How does AI fit into this picture? Will its exorbitant computational cost (and carbon footprint) lead to centralisation, or could it be used to enable smaller enterprises who cannot afford the army of engineering, design, technical and legal experts in a giant corporation. Many years ago, at a Tiree Tech Wave, we imagined a world of digital artisans, individuals or small companies that use digital fabrication to repair, modify or upgrade consumer goods enhancing local community and culture and increasing sustainability. Can we design AI that makes this possible?
Slides
Notes
The talk was in three main parts:
- digital changes everything
- digital changes nothing
- the digital artisan
We'll look at each of these in a little more detail below. However, there is also a big underlying question. If we look back at past technological innovations, we see that, unhindered, technology always favours the big and powerful. Can radical technology instead empower the small and powerless?
This talk considered at one facet of this. It is not so much an account of existing finished work, more an opening up ways of thinking and research challenges.
Digital changes everything
When we look at the impact of AI and digital technology, we can consider two aspects:
- what AI does
- Every technology has direct effects of use; some good, some less so.
- how AI shapes society
- In addition, some technologies fundamentally reshape society. Cars are one example; even if you do not own or use a car the structure of cities, the distances to shops are all shaped by the car. Digital technology
On the first, what AI does, there are clearly good things that AI and digital technology have achieved, indeed the 2024 Nobel Prize for Chemistry was awarded to the developers of AlphaFold. However, there is a dark side; for example, back in the early 1990s I wrote an article warning of the dangers of ethnic, gender and scio-economic bias in black-box machine learning ... a prediction that sadly has come to pass.
On the second, digital technology inevitably leads to what economists call 'netwrok effect', the way that the value an application to you may depend on who else is using it, a positive feedback loop leading to runaway 'winner takes all' emergent monopolies, which undermine the fundamental basis of market economics. Furthermore, the 'new AI' based on deep learning and related methods depends crucially on big data and big computation, further accelerating these feedback loops with impacts on the environment and democracy.
For more about this aspect see the notes of a previous talk "AI Changes Everything".
Digital changes nothing
The South Wales valleys lie north of Cardiff, in the 19th century delivering the coal that powered transport and industry across the world. The rivers run from the mountains as one would expect ... except that the rock strata beneath is a syncline, the rock layers forma bowel-like structure. Given this one might expect a lake on the middle or a fan-like pattern of rivers running inwards before making their way as a single huge river to the sea. The river pattern does not match the geology. The reason is it is, what geographers call, a superimposed drainage pattern: the rivers were there before the syncline formed and as the rock moved simply cut their paths deeper in defiance of the changes below.
It is often the same with digital technology. The digital geology is shifting beneath our feet, but our social and industrial topography remains rooted in the physical and organisational constraints of the 19th century. We create digital solution here and there that replace or enhance a job or process, but the larger scale organisations and structures remain the same.
Deep digitality encourages you to take a more radical stance, reimagining how the world could be completely different. What would industry be like if the silicon revolution had come before the steel revolution? What would banking be like if digital technology has been available in the tome of the Medicis?
One example of this would be if washing machine makers were transformed into IP companies, rater like NVIDIA, licensing out physical manufacture. Some parts, such as the drums, might still be better made in large centralised factories, but it would become far easier for the local repair person to download the schematics of, say, a worn out water pump, pay a small licence fee, and print a fresh one using the 3D printer in the back of their van. This would enable not just repair but replacing the facia to match a new kitchen, or upgrading the interface. Better washing machines for the individual and less waste for the environment.
For more about this deep digitality see my ACM Interactions article "Deep digitality: fate, fiat, and foundry" and the Deep Digitality microsite.
The digital artisan
When we forst thought about the printed washing machine at a Tiree Tech Wave, we called the washing machie repairer/tinkerer a digital artisan. Individual workers or small local companies that are able to repair, upgrade and maybe create new appliances aided by digital fabrication and other digital technology. It is both a concept for a different future, but also a description of things we can see already starting to happen. Let's see some examples.
Steve Gill owns a boat, but some parts are hard to obtain. When this happens he will create a CAD design and then send it to a mail order company with high-quality 3D printers and get the part delivered within days. Steve Harris posted on Facebook about a small tool that he needed to repair a mixer tap. The tool would have needed to be shipped form China, but instead he downloaded plans and printed it in plant-based plastic on his own 3D printer. On Tiree the Maritime Trust had repaired old boats and needed a shed to put them. A team of architects built a shed, called the Noust, as part fo a one week summer school. The smooth shapes of the main roof support beams were designed to reflect the curves of a boat and were constructed form digital plans by using a 2.4mx1.2m CNC to cut out large plywood pieces (to give precise shape) which were the sandwiched around softwood (to give strength). the whole shed was then hand constructed like a giant IKEA flatpack!
These are all inspiring examples, but the people involved have unusual levels of digital and construction expertise. For example, Steve Gill is a product designer and co-authored the TouchIT book on physical and digital design. Could AI help bring down the barriers, not necessarily so that everyone could do this kind of things, but at least to the extent that one could imagine each plumber, electrician, or local engineering firm to be digital artisans?
Creating technology for individuals and small companies is harder than for large factories. There is greater heterogeneity, less control and crucially smaller budgets, meaning that design has to be smarter, not simply throwing computation and money at problems. This offers lots of opportunities for interesting and worthwhile research:
- design
- In the examples we've seen the use of CAD and 3D printing. This often requires expert knowledge such as tolerances and how they differ between materials – a classic expert systems opportunity. We've also seen in the Noust how traditional crafts skill and digital fabrication can be combined. We've seen how 3D printing can be used to make tools as well as being part of a finished object. In addition most crafts include pattern pieces or jigs that are used as part of the construction process, these are ideal for digital fabrication as they ofren do not need to be made of high-resilience materials. Design tools could take a step further offering advice on typical wear patterns and ways to avoid this, fr example, the way that built-in kitchen cupboards often start to sag at the hinges and catch after a few years.
- day-to-day business
- Often small businesses, digital and physical, find themselves spending more time on paperwork than their actual business. This does not scale linearly and large businesses can dedicate staff to finance or understanding new rules and regulations. There are already digital finance applications that help and also the Internet has enabled virtual (human) assistants. Again this sounds like a place where AI could help, especially when combined with remote heman expertise.
- platforms
- This all seems to call out for some form of platforms that maybe specific to types of artisans (electrical goods, textiles, carpentry, etc.). However, what is needed is the opposite of what many gig economy platforms typically privide, which typically avoid or bypass the local health and safety and other legal requirements entirely leaving them to be followed (or not) by the platform users.
Summary
- AI changes everything
- runaway feedback – bigger and bigger
- rewriting market economics
- AI changes nothing
- more of the same, just digitised?
- deep digitality – reimagining possible presents
- the digital artisan
- smaller and smaller
- but hard ... so opportunities for research!
Parts of this work have been supported by the HORIZON Europe projects TANGO - Grant Agreement n. 101120763. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the European Health and Digital Executive Agency (HaDEA). Neither the European Union nor the granting authority can be held responsible for them.
|
