Chair, 695th Lord Mayor's Ethical AI Initiative. Personal views only. Blog on science & religion etc. Work in Strategy & Search. I only follow people I’ve met.
#unethicalOptimizationPrinciple paper now out in @royalsociety Open Science. Thanks to brilliant co-authors Heather Battey, Anthony Davison & Robert Mackay.
@DameWendyDBE gave us quote for press: "This is a really important paper. It shows that…
(1/n)
https://t.co/XEwCR7LV0Y
As Dame Antonia Romeo walked into a packed lecture theatre at the London School of Economics, a ripple of surprise moved across the room...
https://t.co/Ta09acG5EV
It’s commendable to rise from the ranks - but despicable to lie about your CV. I wonder what attracted their Defence Spokesman to “Reform UK”? And how much they’re paying him?
This is very important work! We showed in June that AI destabilises Cybersecurity & so the only satisfactory end-state is formally verified systems. Achieving formal verification of the Signal protocols is a massive step forward on this route.
Announcing a major milestone achieved by my team @GoogleDeepMind: we have proven key security properties of the machine code underlying the Signal messaging protocol. This is the first time the actual production code of a widely deployed protocol has been formally verified – by any method, with or without AI – and at this depth.
More than a research breakthrough, it’s a prime example of how frontier AI can be a force multiplier to counter the speed and scale of modern cyber threats.
Formal verification offers a higher standard of security; however, it’s extremely complex and time consuming, which is why in many cases it’s considered to be impractical. Similar efforts could take teams decades of combined effort. We did it in a month’s time.
Using AlphaProof Code, Google DeepMind's new agentic framework for formal verification that runs on LLMs including Gemini Carbon, we verified the Signal protocol down to the actual code running on devices. This milestone lays the groundwork for a future in which vital systems, from medical device firmware to power grid software, ship with mathematical proof that their protections work.
Reaching that future will take continued research from our community. In that spirit, we're open-sourcing our formal specifications and verification artifacts so others can build on this work.
On a personal note, this moment is of deep significance to me. I started my own research journey more than two decades ago as an intern working on checking the consistency of complex software systems. Seeing this grand challenge made practical with modern AI is extraordinary.
I'm extremely proud of my team members: Swarat Chaudhuri, Andrew Ferraiuolo, Anton Kovsharov, Tej Chajed, Sam Schlesinger, Karn Seth, Jonathan Katz, and Charles Sutton for taking on this grand challenge and delivering, and grateful to the Signal team for their partnership.
https://t.co/kN1eRYQ8fF
@pushmeet@GoogleDeepMind This is very important work! We showed in June that AI destabilises Cybersecurity & so the only satisfactory end-state is formally verified systems. Achieving formal verification of the Signal protocols is a massive step forward on this route.
@montie “That should be the aim if Reform’s leaders are at all serious about building a majority party” But they obviously are neither serious, nor capable of doing so. No-one competent will work for Farage.
An absolute gem of a clip of a young Frank Sinatra singing opera - the famous "Là ci darem la mano" duet from Mozart's Don Giovanni - with a 24-year-old Kathryn Grayson in “It Happened in Brooklyn” (1947). Grayson went on to become MGM’s resident soprano, but Sinatra’s phrasing and breath control is infinitely better, and timbre of his baritone is gorgeous.
VG to see @WilliamJHague@ChiOnwurah@TimClementJone1 & many others at @oiioxford 25th anniversary celebration at House of Lords. Told current OII Director how @WCITLivery Lords colloquium in 1997 on Ethical & Spiritual implication of Internet led to Steve Shirley endowing it.
@DanNeidle You'll have seen that @Jemima_Khan has stated that "The harassment injunction was upheld by the High Court and Court of Appeal on several occasions."
Might it be useful to understand what these "several occasions" were?
@BenGoldsmith@Jemima_Khan It would have been perfectly possible to get an anti-harassment injunction that didn’t have such draconian terms. This one was grotesque overreach.
@ChelFulhamBen This is laudable, but it’s incredibly difficult to combine an open system with adequate cybersecurity - not that cybersecurity is at all easy even in a closed system.🙁🙁