We are happy to once again share an update about our work this past half year, and our current public plans for new work. (See previous Mid-2025 Update) We do not have a substantive update about funding, and are seeking some additional funding to receive matching funds from Survival and Flourishing; without this, our basic work through most of 2026 is fully funded.
- We have announced a new annual AI Safety Research Prize for research done in Israel, jointly with the Israeli Association for Artificial Intelligence. Submissions are due by December 31 AOE. We are grateful to our advisors, Reuth Mirsky, Sarah Keren, and Scott Aaronson, who will be doing a final review of papers for the prize.
We have gotten the Israeli Ministry of Health to commit to a regulation requiring iodine fortification of table salt, and they have said the regulation has been sent to the regulatory authority, but they have been working on how to announce this publicly, how to monitor it, the impact on prices, and other similar topics, which have taken additional time. They are optimistic that this will be completed soon, but do not have details on exactly when it will be announced and implemented.
AI Policy and Standards
- The ISO draft on “Guidance for Human Oversight of AI Systems” has been sent to states for the “DIS Ballot” (Process explanation) which will take until February, at which point in theory the standard can be finalized. We are hoping to push for some changes to ensure that the standard incorporates important minor phrasing changes to ensure human oversight requirements are clear. (The “target date” for the final publication is late 2026, but if technical changes are required, it could require an additional “FDIS” ballot, which could extend the timeline.)
- Other ISO standards work is ongoing, including on Human-Machine teaming, on Guidance on addressing risks in generative AI systems, and on terminology.
- David is helping co-organize a consensus AI standards project with a long list of organizations. This is being launched at AAAI the week of Jan 20th. Irving Torres is leading the coalition management, and Andrea Loehr is leading the systematic review. David is primarily assisting with methodology, connections, and fundraising.
Our paper on AI Oversight (with Aiden Homewood, at GovAI) was accepted to AIGOV@AAAI2026, and will be presented there on Jan 27th. The updated paper is available on ArXiV.
Biorisk and AIxBio
- We are starting work on possible changes to Israeli biosecurity regulation, alongside the national DURC-oversight board, with advice from international partners. David will be presenting about this at a national conference on biorisk on December 30th.
David’s RAND report on future AIxBio Pathogen Design was published. He presented that work the morning of December 10th to the Biological Weapons Convention at a side event organized by RAND.
Other
- David’s paper (mentioned in the previous update) on Semiosis and AI Safety was accepted to Philosophy and Technology, and is forthcoming, eventually.
- Vanessa’s research with several others has continued, with additional progress; see CORAL for details.
