The latest EU AI Act Newsletter:

𝗧𝗵𝗲 𝗘𝗨 𝗔𝗜 𝗔𝗰𝘁 𝗡𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿 #𝟴𝟲: 𝗖𝗼𝗻𝗰𝗲𝗿𝗻𝘀 𝗔𝗿𝗼𝘂𝗻𝗱 𝗚𝗣𝗧-𝟱 𝗖𝗼𝗺𝗽𝗹𝗶𝗮𝗻𝗰𝗲 was just sent to 47,000+ subscribers worldwide:

International Association for Safe & Ethical AI conference earlier this year was excellent. There's another opportunity to attend/present at the next one in Paris on 𝟮𝟰-𝟮𝟲 𝗙𝗲𝗯𝗿𝘂𝗮𝗿𝘆 𝟮𝟬𝟮𝟲!

Europe won't compromise on its right to regulate American Big Tech companies despite threats by President Donald Trump, European Commission President Ursula von der Leyen said today.

Cycling is more than just exercise. It means a cleaner planet, thriving cities, and a stronger economy. EU cycling in numbers: ✔️ 92,000+ km of EuroVelo cycling routes across Europe ✔️ Nearly 10 million bikes produced in the EU every year ✔️ 1.35 million jobs supported by the EU cycling sector

Read my article at @euractiv.com: www.euractiv.com/section/tech...

If xAI cared about AI safety he could have done this research there. But it doesn’t, so he left

In 1865, a catastrophic steam boiler explosion paved the way for independent safety inspections. Today we face a far more complex challenge in AI, but the lessons of the past offer a compelling guide on how to prevent disasters, Santeri Koivula and Alejandro Tlai write.

FLI Is hiring for an 𝗔𝗜 𝗦𝗮𝗳𝗲𝘁𝘆 𝗜𝗻𝘃𝗲𝘀𝘁𝗶𝗴𝗮𝘁𝗼𝗿! The application deadline is 𝟰 𝗦𝗲𝗽𝘁𝗲𝗺𝗯𝗲𝗿 𝟮𝟬𝟮𝟱. The salary range is between $𝟵𝟬,𝟬𝟬𝟬 𝗮𝗻𝗱 $𝟭𝟱𝟬,𝟬𝟬𝟬 per year.

Judging by my email auto-responses, I gather that the entire Europe is on holiday. However, EU AI policy doesn’t stop then, does it? Enter 𝗧𝗵𝗲 𝗘𝗨 𝗔𝗜 𝗔𝗰𝘁 𝗡𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿 #𝟴𝟯: 𝗚𝗣𝗔𝗜 𝗥𝘂𝗹𝗲𝘀 𝗡𝗼𝘄 𝗔𝗽𝗽𝗹𝘆.

Cancer is becoming much less deadly—and scientists’ fast-improving understanding of the disease holds out the prospect of further advances for years to come

The AI Office has launched a €𝟵 𝗺𝗶𝗹𝗹𝗶𝗼𝗻 𝘁𝗲𝗻𝗱𝗲𝗿 to procure technical support for the enforcement of the AI Act, focusing on assessing and monitoring systemic risks posed by General-Purpose AI at EU level. Deadline: 𝟮𝟱 𝗔𝘂𝗴𝘂𝘀𝘁 𝟮𝟬𝟮𝟱.

Laura Caroli and I wrote an op-ed in @techpolicypress.bsky.social: 𝗘𝗨 𝗥𝘂𝗹𝗲𝘀 𝗳𝗼𝗿 𝗚𝗲𝗻𝗲𝗿𝗮𝗹 𝗣𝘂𝗿𝗽𝗼𝘀𝗲 𝗔𝗜 𝗠𝗼𝗱𝗲𝗹 𝗗𝗲𝘃𝗲𝗹𝗼𝗽𝗲𝗿𝘀 𝗔𝗿𝗲 𝗥𝗲𝗮𝗱𝘆, 𝗗𝗲𝘀𝗽𝗶𝘁𝗲 𝗪𝗵𝗮𝘁 𝗜𝗻𝗱𝘂𝘀𝘁𝗿𝘆 𝗦𝗮𝘆𝘀. The compliance tools are ready, the timelines are reasonable, and the rules reflect what many companies already do.

If you haven't subscribed yet to the EU AI Act newsletter, do so here: 😉 artificialintelligenceact.substack.com?utm_source=s...

Proton is so good 🙏

The latest EU AI Act Newsletter:

𝗕𝗶𝗴 𝗻𝗲𝘄𝘀: The European Commission is setting up a 𝘀𝗰𝗶𝗲𝗻𝘁𝗶𝗳𝗶𝗰 𝗽𝗮𝗻𝗲𝗹 𝗼𝗳 𝗶𝗻𝗱𝗲𝗽𝗲𝗻𝗱𝗲𝗻𝘁 𝗲𝘅𝗽𝗲𝗿𝘁𝘀 to support the implementation and enforcement of the AI Act. Applications are open until 𝟭𝟰 𝗦𝗲𝗽𝘁𝗲𝗺𝗯𝗲𝗿.

𝗧𝗵𝗲 𝗘𝗨 𝗔𝗜 𝗔𝗰𝘁 𝗡𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿 #𝟳𝟵: 𝗖𝗼𝗻𝘀𝘂𝗹𝘁𝗮𝘁𝗶𝗼𝗻 𝗼𝗻 𝗛𝗶𝗴𝗵-𝗥𝗶𝘀𝗸 𝗔𝗜 was just published!

The EU could be a promising option, a lot of funding available: commission.europa.eu/topics/resea...

𝗧𝗵𝗲 𝗘𝗨 𝗔𝗜 𝗔𝗰𝘁 𝗡𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿 #𝟳𝟴: 𝗖𝘂𝘁𝘁𝗶𝗻𝗴 𝗥𝗲𝗱 𝗧𝗮𝗽𝗲 was just published! We are looking for 𝟮-𝗺𝗶𝗻𝘂𝘁𝗲 𝗳𝗲𝗲𝗱𝗯𝗮𝗰𝗸 on the AI Act website, so that we can build the resources most helpful to you. We have produced many of the resources on our website in direct response to user feedback.

Thanks, @ristouuk.bsky.social 🙏

Together with Sten Tamkivi, partner at Plural, a leading European early-stage venture fund, we argue that 𝘁𝗵𝗲 𝗘𝗨 𝘀𝗵𝗼𝘂𝗹𝗱 𝗰𝘂𝘁 𝗮𝗰𝘁𝘂𝗮𝗹 𝗿𝗲𝗱 𝘁𝗮𝗽𝗲, 𝗻𝗼𝘁 𝗔𝗜 𝘀𝗮𝗳𝗲𝗴𝘂𝗮𝗿𝗱𝘀 in an op-ed in Fortune.

𝗧𝗵𝗲 𝗘𝗨 𝗔𝗜 𝗔𝗰𝘁 𝗡𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿 #𝟳𝟳: 𝗔𝗜 𝗢𝗳𝗳𝗶𝗰𝗲 𝗧𝗲𝗻𝗱𝗲𝗿 has been published! 𝘛𝘩𝘦 𝘈𝘐 𝘖𝘧𝘧𝘪𝘤𝘦 𝘸𝘪𝘭𝘭 𝘴𝘰𝘰𝘯 𝘣𝘦 𝘭𝘰𝘰𝘬𝘪𝘯𝘨 𝘧𝘰𝘳 𝘵𝘩𝘪𝘳𝘥-𝘱𝘢𝘳𝘵𝘺 𝘤𝘰𝘯𝘵𝘳𝘢𝘤𝘵𝘰𝘳𝘴 𝘵𝘰 𝘴𝘶𝘱𝘱𝘰𝘳𝘵 𝘵𝘩𝘦 𝘮𝘰𝘯𝘪𝘵𝘰𝘳𝘪𝘯𝘨 𝘰𝘧 𝘤𝘰𝘮𝘱𝘭𝘪𝘢𝘯𝘤𝘦 𝘢𝘯𝘥 𝘳𝘪𝘴𝘬 𝘢𝘴𝘴𝘦𝘴𝘴𝘮𝘦𝘯𝘵 𝘰𝘧 𝘨𝘦𝘯𝘦𝘳𝘢𝘭-𝘱𝘶𝘳𝘱𝘰𝘴𝘦 𝘈𝘐 𝘮𝘰𝘥𝘦𝘭𝘴.

My colleague Santeri Koivula at Future of Life Institute (FLI) has put together a new resource 𝗔𝗜 𝗥𝗲𝗴𝘂𝗹𝗮𝘁𝗼𝗿𝘆 𝗦𝗮𝗻𝗱𝗯𝗼𝘅 𝗔𝗽𝗽𝗿𝗼𝗮𝗰𝗵𝗲𝘀: 𝗘𝗨 𝗠𝗲𝗺𝗯𝗲𝗿 𝗦𝘁𝗮𝘁𝗲 𝗢𝘃𝗲𝗿𝘃𝗶𝗲𝘄. This has been one of the most in-demand additions to our EU AI Act website.

The AI Office announced a tender titled 𝗔𝗿𝘁𝗶𝗳𝗶𝗰𝗶𝗮𝗹 𝗜𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 𝗔𝗰𝘁: 𝗧𝗲𝗰𝗵𝗻𝗶𝗰𝗮𝗹 𝗔𝘀𝘀𝗶𝘀𝘁𝗮𝗻𝗰𝗲 𝗳𝗼𝗿 𝗔𝗜 𝗦𝗮𝗳𝗲𝘁𝘆 with an estimated total value over €𝟵 𝗺𝗶𝗹𝗹𝗶𝗼𝗻:

I enjoyed the discussion about AI's existential risks with Reid Blackman and Torben Swoboda on the Ethical Machines podcast:

𝗧𝗵𝗲 𝗘𝗨 𝗔𝗜 𝗔𝗰𝘁 𝗡𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿 #𝟳𝟱: 𝗔𝗜 𝗖𝗼𝗻𝘁𝗶𝗻𝗲𝗻𝘁 𝗔𝗰𝘁𝗶𝗼𝗻 𝗣𝗹𝗮𝗻 has been published! 𝘛𝘩𝘦 𝘌𝘶𝘳𝘰𝘱𝘦𝘢𝘯 𝘊𝘰𝘮𝘮𝘪𝘴𝘴𝘪𝘰𝘯 𝘩𝘢𝘴 𝘪𝘯𝘵𝘳𝘰𝘥𝘶𝘤𝘦𝘥 𝘵𝘩𝘦 𝘈𝘐 𝘊𝘰𝘯𝘵𝘪𝘯𝘦𝘯𝘵 𝘈𝘤𝘵𝘪𝘰𝘯 𝘗𝘭𝘢𝘯 𝘵𝘰 𝘭𝘦𝘷𝘦𝘳𝘢𝘨𝘦 𝘌𝘜 𝘴𝘵𝘳𝘦𝘯𝘨𝘵𝘩𝘴 𝘭𝘪𝘬𝘦 𝘵𝘢𝘭𝘦𝘯𝘵 𝘢𝘯𝘥 𝘴𝘵𝘳𝘰𝘯𝘨 𝘵𝘳𝘢𝘥𝘪𝘵𝘪𝘰𝘯𝘢𝘭 𝘪𝘯𝘥𝘶𝘴𝘵𝘳𝘪𝘦𝘴 𝘢𝘴 𝘈𝘐 𝘢𝘤𝘤𝘦𝘭𝘦𝘳𝘢𝘵𝘰𝘳𝘴.

𝗧𝗵𝗲 𝗘𝗨 𝗔𝗜 𝗔𝗰𝘁 𝗡𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿 #𝟳𝟰: 𝗛𝘂𝗺𝗮𝗻 𝗥𝗶𝗴𝗵𝘁𝘀 𝗔𝗿𝗲 𝗡𝗼𝘁 𝗢𝗽𝘁𝗶𝗼𝗻𝗮𝗹 was just emailed to 42,000+ subscribers. 𝐴𝑟𝑐ℎ𝑖𝑡𝑒𝑐𝑡𝑠 𝑜𝑓 𝑡ℎ𝑒 𝐴𝐼 𝐴𝑐𝑡 ℎ𝑎𝑣𝑒 𝑢𝑟𝑔𝑒𝑑 𝐵𝑟𝑢𝑠𝑠𝑒𝑙𝑠 𝑡𝑜 ℎ𝑎𝑙𝑡 "𝑑𝑎𝑛𝑔𝑒𝑟𝑜𝑢𝑠" 𝑤𝑎𝑡𝑒𝑟𝑖𝑛𝑔 𝑑𝑜𝑤𝑛 𝑜𝑓 𝑟𝑢𝑙𝑒𝑠 𝑡ℎ𝑎𝑡 𝑤𝑜𝑢𝑙𝑑 𝑒𝑥𝑒𝑚𝑝𝑡 𝑚𝑎𝑗𝑜𝑟 𝑈𝑆 𝑡𝑒𝑐ℎ 𝑐𝑜𝑚𝑝𝑎𝑛𝑖𝑒𝑠 𝑙𝑖𝑘𝑒 𝑂𝑝𝑒𝑛𝐴𝐼 𝑎𝑛𝑑 𝐺𝑜𝑜𝑔𝑙𝑒 𝑓𝑟𝑜𝑚 𝑘𝑒𝑦 𝑟𝑒𝑔𝑢𝑙𝑎𝑡𝑜𝑟𝑦 𝑟𝑒𝑞𝑢𝑖𝑟𝑒𝑚𝑒𝑛𝑡𝑠.

I have excellent news: our KU Leuven 𝗜𝗻𝘁𝗲𝗿𝗻𝗮𝘁𝗶𝗼𝗻𝗮𝗹 𝗖𝗼𝗻𝗳𝗲𝗿𝗲𝗻𝗰𝗲 𝗼𝗻 𝗟𝗮𝗿𝗴𝗲-𝗦𝗰𝗮𝗹𝗲 𝗔𝗜 𝗥𝗶𝘀𝗸𝘀 is now open for registration for the broader public! We received a lot of interesting talk submissions and we've made our selection of the best ones. It will take place on 𝟮𝟲-𝟮𝟴 𝗠𝗮𝘆 𝗶𝗻 𝗟𝗲𝘂𝘃𝗲𝗻, 𝗕𝗲𝗹𝗴𝗶𝘂𝗺.

𝗧𝗵𝗲 𝗘𝗨 𝗔𝗜 𝗔𝗰𝘁 𝗡𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿 #𝟳𝟯: 𝗦𝗰𝗶𝗲𝗻𝘁𝗶𝗳𝗶𝗰 𝗣𝗮𝗻𝗲𝗹 𝗥𝘂𝗹𝗲𝘀 is out! 𝘛𝘩𝘦 𝘊𝘰𝘮𝘮𝘪𝘴𝘴𝘪𝘰𝘯 𝘩𝘢𝘴 𝘦𝘴𝘵𝘢𝘣𝘭𝘪𝘴𝘩𝘦𝘥 𝘳𝘶𝘭𝘦𝘴 𝘧𝘰𝘳 𝘢 𝘯𝘦𝘸 𝘴𝘤𝘪𝘦𝘯𝘵𝘪𝘧𝘪𝘤 𝘢𝘥𝘷𝘪𝘴𝘰𝘳𝘺 𝘨𝘳𝘰𝘶𝘱 𝘰𝘧 𝘪𝘯𝘥𝘦𝘱𝘦𝘯𝘥𝘦𝘯𝘵 𝘈𝘐 𝘦𝘹𝘱𝘦𝘳𝘵𝘴.

I'm happy to share a new policy brief titled 𝗟𝗼𝗼𝗸𝗶𝗻𝗴 𝗔𝗵𝗲𝗮𝗱: 𝗦𝘆𝗻𝗲𝗿𝗴𝗶𝗲𝘀 𝗕𝗲𝘁𝘄𝗲𝗲𝗻 𝘁𝗵𝗲 𝗘𝗨 𝗔𝗜 𝗢𝗳𝗳𝗶𝗰𝗲 𝗮𝗻𝗱 𝗨𝗞 𝗔𝗜𝗦𝗜 published on the Oxford Martin School's website. This is work carried out with Lara Thurnherr, Tekla Emborg, Marta Ziosi, Isabella Wilkinson, Morgan Simpson, Renan Araujo, and Charles Martinet.

PYMNTS wrote an article about the Code of Practice for general-purpose AI models with some quotes from me:

I'm happy to share that our paper "𝗢𝗽𝗲𝗿𝗮𝘁𝗶𝗼𝗻𝗮𝗹𝗶𝘀𝗶𝗻𝗴 𝘁𝗵𝗲 𝗱𝗲𝗳𝗶𝗻𝗶𝘁𝗶𝗼𝗻 𝗼𝗳 𝗴𝗲𝗻𝗲𝗿𝗮𝗹-𝗽𝘂𝗿𝗽𝗼𝘀𝗲 𝗔𝗜: 𝗮𝘀𝘀𝗲𝘀𝘀𝗶𝗻𝗴 𝗳𝗼𝘂𝗿 𝗮𝗽𝗽𝗿𝗼𝗮𝗰𝗵𝗲𝘀" with Carlos Ignacio Gutierrez and Alex Tamkin has been published in Law, Innovation and Technology.

𝗧𝗵𝗲 𝗘𝗨 𝗔𝗜 𝗔𝗰𝘁 𝗡𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿 #𝟳𝟮: 𝗘𝘂𝗿𝗼𝗽𝗲'𝘀 𝗖𝗼𝗺𝗽𝗲𝘁𝗶𝘁𝗶𝘃𝗲𝗻𝗲𝘀𝘀 is out now! You're welcome to share it with your colleagues and whoever else might be interested. 𝘛𝘩𝘦 𝘌𝘶𝘳𝘰𝘱𝘦𝘢𝘯 𝘛𝘦𝘤𝘩 𝘈𝘭𝘭𝘪𝘢𝘯𝘤𝘦 𝘱𝘢𝘱𝘦𝘳 "𝘌𝘯𝘴𝘶𝘳𝘪𝘯𝘨 𝘌𝘶𝘳𝘰𝘱𝘦'𝘴 𝘊𝘰𝘮𝘱𝘦𝘵𝘪𝘵𝘪𝘷𝘦𝘯𝘦𝘴𝘴" 𝘰𝘶𝘵𝘭𝘪𝘯𝘦𝘴 𝘴𝘵𝘦𝘱𝘴 𝘧𝘰𝘳 𝘌𝘶𝘳𝘰𝘱𝘦𝘢𝘯 𝘤𝘰𝘮𝘱𝘢𝘯𝘪𝘦𝘴 𝘵𝘰 𝘳𝘦𝘮𝘢𝘪𝘯 𝘨𝘭𝘰𝘣𝘢𝘭𝘭𝘺 𝘤𝘰𝘮𝘱𝘦𝘵𝘪𝘵𝘪𝘷𝘦 𝘪𝘯 𝘈𝘐.

It was fun to appear on the Age of AI podcast with Torben Swoboda. We discussed several key topics: the challenges of defining AI risk, the debate over acceptable thresholds, common critiques of the existential risk perspective, and the feasibility of global AI governance.

I'm excited to share that I’m joining the Stanford Digital Economy Lab for a research visit. I have been very impressed by the intellectual environment at Stanford University every time I’ve visited. I’m very grateful for this opportunity.

GenAI as a business "costs are also so severe that .. paying customers lose .. companies money. .. the most successful company .. appears to have no way to stop burning money... .. only one real company in this industry, OpenAI, and it is ... not a real business." www.wheresyoured.at/wheres-the-m...

My colleague Tekla Emborg has put together a new summary of 𝘄𝗵𝗮𝘁 𝘁𝗵𝗲 𝗔𝗜 𝗔𝗰𝘁 𝗺𝗲𝗮𝗻𝘀 𝗳𝗼𝗿 𝘀𝗺𝗮𝗹𝗹 𝗮𝗻𝗱 𝗺𝗲𝗱𝗶𝘂𝗺 𝗯𝘂𝘀𝗶𝗻𝗲𝘀𝘀𝗲𝘀, on the EU AI Act website:

𝗧𝗵𝗲 𝗘𝗨 𝗔𝗜 𝗔𝗰𝘁 𝗡𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿 #𝟳𝟭: 𝗪𝗲𝗮𝗹𝘁𝗵 𝗼𝗳 𝗖𝗼𝗺𝗺𝗶𝘀𝘀𝗶𝗼𝗻 𝗚𝘂𝗶𝗱𝗲𝗹𝗶𝗻𝗲𝘀 has been published. This will land in 40,000+ subscribers' inboxes and celebrate the newsletter's 3-year anniversary. I hope you find it useful.

📘 Want to experience working for the EU? The Blue Book traineeship applications are OPEN! 💶 5-month paid traineeship 📅 Starts in October 2025 🎓 You can apply with any Bachelor’s degree

Three fallacies of AI: 1. “That the purpose of AI is efficiency and scale” 2. “That AI requires a tradeoff – between safety and progress, between competition and collaboration, and between rights and innovation” 3. “It is not inevitable that AI will lead to great public benefits”

READ EVERY LETTER: America is on the cusp of competitive authoritarianism. The Trump administration has already begun to weaponize state institutions and deploy them against opponents. The Constitution alone cannot save U.S. democracy ↘️ www.foreignaffairs.com/united-state...

New video 📽️ from me: I had the privilege of being a keynote speaker at the International Association for Safe & Ethical AI. I focused on solutions and ideas for moving forward. They made it public, and here it is. Enjoy, and hope it's helpful! 🤗 oecdtv.webtv-solution.com/embed-or-en-...

@futureoflife.org AI Action Summit Lead Imane (Ima) Bello has put together a useful guide for policymakers to navigate the AI Action Summit in Paris on 10-11 February:

𝗧𝗵𝗲 𝗘𝗨 𝗔𝗜 𝗔𝗰𝘁 𝗡𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿 #𝟳𝟬: 𝗙𝗶𝗿𝘀𝘁 𝗠𝗲𝗮𝘀𝘂𝗿𝗲𝘀 𝗧𝗮𝗸𝗲 𝗘𝗳𝗳𝗲𝗰𝘁 was just published. You're welcome to share it with your friends and colleagues. Thanks for reading!

What did you do when you were 15 years old? I know what I definitely didn’t do – I didn’t record podcasts with a US congresswoman, MIT academics or an NBA champion, nor did I think much about regulation. That's exactly what Charlie Zeitlin has been doing for a couple of years now.

Today, we are publishing the first-ever International AI Safety Report, backed by 30 countries and the OECD, UN, and EU. It summarises the state of the science on AI capabilities and risks, and how to mitigate those risks. 🧵 Full Report: assets.publishing.service.gov.uk/media/679a0c... 1/21

One thing that I absolutely loved when I studied at the @lsephilosophy.bsky.social was all of the great events that @lseevents.bsky.social organized. Just consistent incredible events for everyone. Such an amazing resource. Luckily we can also watch the recordings online:

The AI executive order revoked and more scaling, not an amazing start for AI safety this year.

I really like @yoshuabengio.bsky.social's humility in the recent Machine Learning Street Talk episode on AI safety, whether it's on probability of doom, different risk scenarios, AGI timeline, or other topics: