1/ a thread about the changes we’ve made to the AI Safety Institute – we are now the AI Security Institute, we’re setting up a new team to research the risk of criminal misuse and we’re deepening our existing partnerships with the UK national security community. 2/ From the outset our focus has always been on the most serious risks that could come from AI. When we launched AISI we said it would focus on chemical and biological risks, cyber, loss of control, safegaurds etc. Our evaluations have been squarely focused on security risks.

1/ It’s been one year since I was appointed Chair of the UK AI Safety Institute. In this time, we’ve built one of the largest safety evaluation teams globally and are already conducting pre-deployment testing. This is our fourth progress report 2/ Report: . Over the last year, we’ve gone from building a start-up inside Government, to shipping product. We’ve built one of the largest safety evaluation teams globally - with a team of over 30 technical researchers and counting.aisi.gov.uk/work/fourth-pr…

1/ I've just left the final session of the first ever global Summit on AI Safety, chaired by @RishiSunak and @michelledonelan. A thread on how it started vs how it’s going: 2/ How it started: we had 4 goals on safety, 1) build a global consensus on risk, 2) open up models to government testing, 3) partner with other governments in this testing, 4) line up the next summit to go further. How it’s going: 4 wins:

1/ The Taskforce is a start-up inside government, delivering on the mission given to us by the Prime Minister: to build an AI research team that can evaluate risks at the frontier of AI. We are now 18 weeks old and this is our second progress report: gov.uk/government/pub… 2/ The frontier is moving very fast. On the current course, in the first half of 2024, we expect a small
handful of companies to finish training models that could produce another significant jump in
capabilities beyond state-of-the-art in 2023.

1/ 11 weeks ago I agreed to Chair the UK's efforts to accelerate state capacity in AI Safety - measuring and mitigating the risks of frontier models so we can safety capture their opportunities. Here is our first progress report: gov.uk/government/pub…

https://twitter.com/SciTechgovuk/status/1699686882022486145

2/ The Taskforce is a start-up inside government, delivering on the ambitious mission given to us by the Prime Minister. Effective start-ups send regular investor updates, so here is ours: gov.uk/government/pub…

I’m honoured to be appointed as the Chair of the UK's AI Foundation Model Taskforce.

A thread on why I'm doing this and how you might be able to help us.

https://twitter.com/SciTechgovuk/status/1670313211625701376

1/ I wrote in 2018 about how accelerating AI progress would create new geopolitical challenges: ianhogarth.com/blog/2018/6/13…

1/ Notable how three pioneers of deep learning ( recognised in their shared 2018 Turing award) have substantially diverged on how they assess risk from superintelligence: 2/ Yoshua Bengio was one of the leading signatories to the open letter calling on "all AI labs to immediately pause for at least 6 months the training of AI systems more powerful than GPT-4" futureoflife.org/open-letter/pa…

1/ One of the most significant research papers of the last year is the Toolformer paper. It provides evidence that "Language Models [e.g. GPT-4] Can Teach Themselves to Use Tools" arxiv.org/abs/2302.04761 2/ This is why OpenAI's 'plugins' release is so significant - they are taking advantage of this deep insight by connecting GPT-4 to lots of digital tools openai.com/blog/chatgpt-p…

A thread of midwit takes on AGI - give a like to any you've seen in the wild

1/ Really excited to announce that @CarinaNamih mRNA pioneer and founder turned investor is joining us at Plural ianhogarth.com/blog/2023/3/23… 2/ At @pluralplatform we are building a platform for serious founders to put capital, experience and energy behind important missions. Carina is the perfect example of someone who does that

1/ Glad to see major AI labs starting to share their AI Alignment plans. Activity in the last two weeks: 2/ Feb 26: @ConjectureAI shared "their primary safety proposal and research direction...Cognitive Emulation (or “CoEm”). The goal of the CoEm agenda is to build predictably boundable systems" lesswrong.com/posts/ngEvKav9…

1/ A few thoughts on TicketMaster and @taylorswift13. 2/ I co-founded and ran @songkick for 8 years. Songkick was founded in 2007 to create a better fan and artist experience around live music. We were backed by YC, Sequoia etc

1/ The @stateofaireport 2022 is live!

For the 5th year, @nathanbenaich and I compile the most important work in AI research, industry, talent, and politics. Our report is open-access to all. Here are some things that really stood out for me…
stateof.ai 2/ @demishassabis vision of using AI to accelerate scientific discovery continues to gather momentum with breakthrough results spanning mathematics, materials science and nuclear fusion

1/ Excited to announce one of Plural's first investments - Field Energy. The team is building and operating grid scale batteries with a plan to get 1.3 gigawatt hours of battery storage operational across the UK by 2024 bloomberg.com/news/articles/… 2/ As we bring more renewables onto the grid, decommission coal peaker plants and electrify major areas of the economy (heating, transportation) we will see huge increase in volatility across energy markets.

1/ When I was building @Songkick, the most valuable investors I had were people who had built things themselves - Greg McAdoo, Paul Graham and Saul Klein had all been founders and CEOs. They helped me learn faster and were far more committed to helping us succeed 2/ For years I was back and forth between Silicon Valley and London. It was obvious the difference in the quality of the investors. In the Valley investors had often been founders and they carried themselves differently - more informal, and more focused on what actually matters.

1/ The @stateofaireport 2021 is live!

For the 4th year, @nathanbenaich and I compile the most important work in AI research, industry, talent, and politics. Our report is open-access to all. Here are some things that really stood out for me... 2/ The dual-use challenge I raised in 2018 (ianhogarth.com/blog/2018/6/13…) has started to materialise with multiple deployments of AI by militaries - most notably by Israel in Gaza attacks

1/ STATE OF AI 2020 IS HERE! For the 3rd year running, @NathanBenaich and I have tried to compile the most interesting developments in AI. Featuring the biggest research breakthroughs, novel commercial applications and the major political developments. stateof.ai 2/ this year we invited 25 of the most interesting AI start-ups to share novel data for the report. Honoured to have contributions from the likes of @disperseio @graphcoreai @huggingface @labgeni_us @openminedorg @poly_ai @PostEra_AI @RecursionPharma @Tessian @tractable_ai.

1/ When I was at university, my friend @tommustill was fascinated with what we could learn about enigmatic creatures, thought to be the earliest animals that lived over 550 million years ago... 2/ he had been spending time north of Archangel in Russia gathering data on fossils - mapping out the spatial location that different organisms had died in and what that implied for the ecosystems they lived in...

1/ lots of interesting stuff in this Pew Research survey on American attitudes to automation: pewinternet.org/2017/10/04/aut… 2/ "6% of Americans report that they have already been impacted by automation in the form of lost jobs and/or wages"

1/ YouTube Red is interesting. All you can eat content subscription. Unlike much else (maybe Amazon Prime?): musicbusinessworldwide.com/youtube-red-13… 2/ Critical that they maintain their unique and huge depth of music catalogue through the transition, so prob smart forcing consistency

1/ interesting to consider the vested interests of different tech cos in a freemium streaming model (vs 100% paid): recode.net/2015/03/08/uni… 2/ important for Spotify to preserve because they have invested considerable energy in optimising this growth engine and it is humming