Rarely a day goes by without an AI story in the news. Sometimes, there is good news, such as the use of AI to discover a new pharmaceutical, but often more dark, about AI bias or the way it may rob us of jobs, privacy or autonomy. The human impact of AI is of two kinds. First, what AI does directly - systems that we use and can design better or worse. Second, how AI shapes society, the way AI can create mismatches of power between large corporations and nation states, and between organisations and individuals. Recent advances in large-langage models in particular may mean that AI is only in the hands of those who can afford massive computational power and technical expertise. However, there are signs of hope, in particular the way that generative AI might enable niches applications that would otherwise be impossible. In education this may allow personalised tuition, bit also changes what needs to be learnt ... not necessarily digital; the ability of LLMs to generalise may offer ways for minority languages to survive; and in health there is the possibility of personalised medicine, and affordable ways to help well-being and mental health.
Slides
|