Thinking differently about AI and risk
Long before ChatGPT I gave a talk on AI and risk. It turns out that it was more prescient than I ever anticipated!
Back in 2021 I gave a presentation to a National Academies of Science-sponsored symposium series on thinking differently about the risks and benefits of AI.
A lot has happened since then — including an escalation in proposals, hearings, discussions, and initiatives, around artificial risk and safety.
And yet, revisiting the talk, I’m surprised at how prescient it was and how relevant the ideas still are — to the degree that I thought it worth reposting:
At the heart of the talk is a discussion of how to frame novel risks associated with AI, and how the risk innovation framework we have developed here at ASU could be applied to good effect.
Remember that this was well before the widespread use of LLMs and other foundation models. Yet the ideas here are perhaps more relevant than ever in an age of generative AI.
The talk is only 20 minutes long — and even shorter on 1.5x playback! I’d highly recommend spending a few minutes watching it for the insights it provides on thinking differently about AI and risk.