iiinova — where your digital self keeps evolving

Aidan P

@aidanpeppin.bsky.social

Policy at Cohere.com. Formerly @adalovelaceinst.bsky.social. Cat is called Lola.

Aidan P @aidanpeppin.bsky.social · 04/12/2024

Huge thanks to co-authors: @ankareuel.bsky.social, Stephen Casper, @elliotmjones.bsky.social, @agstrait.bsky.social , Usman Anwar, Anurag Agrawal, @sayash.bsky.social, Sanmi Koyejo, Marie Pellat, Rishi Bommasani, @nickfrosst.bsky.social & @sarahooker.bsky.social

080

Aidan P @aidanpeppin.bsky.social · 04/12/2024

We find that current suggests no increased risk from current models, & methods to study AI biorisk are nascent, often speculative, & limited in terms of methodological maturity. More work is needed to develop theoretical models for biorisk & empirical methods to test them.

171

Aidan P @aidanpeppin.bsky.social · 04/12/2024

AI amplifying biorisk has been a major topic in policy & governance work. But does the available evidence match this level of attention? 🦠 ⚠️ Our new paper looks at the science underpinning ideas that AI could increase biorisks. arxiv.org/abs/2412.01946

arxiv.org

The Reality of AI and Biorisk

To accurately and confidently answer the question 'could an AI model or system increase biorisk', it is necessary to have both a sound theoretical threat model for how AI models or systems could incre...

1162

Aidan P @aidanpeppin.bsky.social · 14/11/2024

the most wonderful time of the year is almost upon us when we can start emails with: "let's kick this to the new year"

020