In this talk we'll take a look at the pragmatic reality of AI through a lens that focuses on real risks and ethical considerations that are playing out right now.
Since ChatGPT's explosive launch, there doesn't seem to be a business or startup in the world that isn't desperately seeking a way to inject AI into their product (or at least, into their pitch decks), filled with hyperbolic claims that it is an inevitable future and anyone not onboard is being left behind.
Along the way, AI 'risk' has captured the public imagination in the form of X-risk - the idea that a malevolent, (or perhaps blindly obedient), AI poses an existential risk to the future of humanity either through turning us all into paperclips or simultaneously activating the global nuclear arsenal.
This talk makes the case that AI is neither the saviour nor the doom of humanity, but it does pose some serious questions that need urgent consideration and the solutions to which can most effectively be realised by the product and engineering teams on the ground.
There are two perspectives to this analysis - supply side risk and product side risk.
On the supply side we consider risks and ethical concerns such as data classification worker rights, IP & copyright theft, energy consumption & privacy considerations.
On the product side we consider the impact of embedded biases, the application of AI in regulated products such as the medical or legal domains & the psychological impact of anthromorphising synthetic text generators.
Finally, we'll share references to notable AI researchers advancing the field and links for further reading.
All talks