How Dangerous Is Artificial Intelligence?
Experts in the field of AI are engaged in a serious, and very contentious debate about whether AI poses an "existential risk" to humanity
This week I have been thinking a great deal about a contentious debate among AI experts — researchers who each have decades of experience in the field — debating whether AI poses an “existential risk” to humanity.
Existential risk here means the risk that artificial intelligence could lead to “human extinction or another irreversible global catastrophe.”
The debate was hosted by Munk Debates (@munkdebate), a cool organization that I have just been introduced to, and you can watch the recording on youtube here:
It’s also available as a podcast.
The speakers in favor of what I’ll call the “AI Doom” position were:
• Yoshua Bengio: Full Professor at Université de Montréal, and the Founder and Scientific Director of Mila – Quebec AI Institute (https://yoshuabengio.org) and
• Max Tegmark (@tegmark): Professor doing AI and physics research at MIT as part of the Institute for Artificial Intelligence & Fundamental Interactions and the Center for Brains, Minds and Machines (https://physics.mit.edu/faculty/max-t...)
The speakers against the proposition were
• Melanie Mitchell (@MelMitchell1): Professor at the Santa Fe Institute (https://melaniemitchell.me)
• Yann LeCun (@ylecun): VP & Chief AI Scientist at Meta and Silver Professor at NYU affiliated with the Courant Institute of Mathematical Sciences & the Center for Data Science (http://yann.lecun.com)
This debate resists easy summary. There are lots of separate issues and lots of people, IMHO anyway, talking past each other. I also believe that it is hugely significant. In an effort to understand the various issues, I have started work on a longer wiki article on the LawSnap wiki website, available here: https://lawsnap.mywikis.wiki/wiki/AI_Risk
My hope is to keep building this out to cover — as fairly as I can — the various positions staked out in these debates. Your feedback is much appreciated.