Expert opinions on International Relations

Building safe and reliable AI will require global regulation in the sense of model testing and evaluation.

AI isn’t going away. Barring unprecedented global government coordination, it isn’t even slowing down. The only real decision is whether to centralize power, or disperse it. Do you fear the technology? Or a small number of people controlling the te...

The U.S. and China won’t move in lockstep on AI safety, and forging mutually binding agreements with real teeth is probably not possible right now.

At the same time, the rapid development of frontier AI models poses serious risks to safety and security if not appropriately managed. [...] We therefore welcome the recently launched initiatives that aim to manage the risks related to frontier techn...

Rather than creating entirely new channels, policymakers should focus on ensuring that existing channels can effectively incorporate information related to frontier models and agents. Building on trusted relationships will be faster and more effectiv...

As the frontier models become more capable, these risks have come more sharply into focus. Models being developed will allow hacking on a level we never thought possible. And they will soon be in the hands of adversaries who see the world very differ...

The question that the noble Lord posed was: do we therefore go for a moratorium? That would be highly desirable, but I do not think that it will be possible in the short term.

The Technology Secretary rejected calls to pause the development of artificial intelligence, describing such a move as “a double betrayal” of British talent and British interests.

That is why scientists around the world have been calling for a CERN-like international AI-research agency for nearly a decade. To date, several half-measures in this direction have been launched. But none come close to establishing a well-funded, lo...

But the main regret I have with this talk is that I wish I had prepared a more for the section on a “CERN for AI.” The idea behind such a project is to pool many countries’ and companies’ resources into a single (possibly physically decentralized) ci...

There are many conceivable versions of a “CERN for AI.” But the version that seems politically realistic (a new catch-up lab) probably would not do much for safety, while the versions that would materially improve safety (e.g., pause + merge of all c...

Today, the companies are being built elsewhere. Last year the United States produced 59 notable AI models and China produced 35. France and the United Kingdom produced one each. Along the rest of the AI value chain, the picture is the same: the Unite...

Europe should prioritize “code over concrete.”

Middle powers should combine complementary national assets and pool the capabilities that require scale, while preserving national and commercial control over applications, deployment, language, and sector-specific use. Shared efforts could include m...

The answer isn't convergence for its own sake. We can push too hard on one thing. But if you have one standard, you could risk a lowest common denominator outcome that can cost us the innovation system as well and might have a risk on the national in...

The bills include the AI OVERWATCH Act, the Chip Security Act, and the MATCH Act. Anthropic supports all three bills and is ending its membership in ITI over the group's efforts to block them from passing.

Our membership is aligned that these bills undermine the American tech stack, and as our Thursday letter demonstrates, there is widespread support for this position across the technology and broader business sectors.

A covered developer — Compute threshold: 10^26 FLOP — has trained or initiated the training of a frontier model, a model trained using, or undergoing training intended to use, at least 10^26 floating-point operations of compute. [...] Covered develop...

According to the proposed rule, a covered US person will be required to provide quarterly reports to BIS if the US person “engages in, or plans, within six months, to engage in ‘applicable activities.’” “Applicable activities” include: “Conducting a...

Policymakers do not need to approach each innovation in isolation and should not treat every emerging technology as a first-of-its-kind policy problem.

This means that in 2026-2027 we could end up in one of two starkly different worlds. In the US, multiple companies will definitely have the required millions of chips (at the cost of tens of billions of dollars). The question is whether China will al...

China consistently opposes any practice of overstretching the concept of national security and abusing export controls.

Government should be pressing Washington for published criteria before access is restricted, advance warning and exemptions for allies, as we already handle security of supply in energy and defense. At home, it should require critical infrastructure ...

But even a country that gets its own house in order will still be exposed to risks that cross borders. This is why an international organization will need to be built in parallel. […] A new global organization for AI will need elements of all three a...

The time has come for a more systematic approach to AI regulation — funded by the industry, staffed by world-class technical experts, and answerable to the U.S. government.

While we continue to work toward a world free from both nuclear weapons and autonomous killing machines, we welcome the focus in the ongoing negotiations on ensuring ‘meaningful human control’ over autonomous weapons.

A broad consensus has emerged that, regardless of the system, humans must retain responsibility for decisions to use force.

A total prohibition on all autonomous weapons would likely fail to distinguish between high-risk anti-personnel systems and narrowly bounded defensive systems used against incoming missiles or munitions.

Any serious European AI initiative must include the United Kingdom. DeepMind, based in London, is one of the world's three or four leading AI research laboratories. ARM's chip architecture underpins a significant share of global computing. The UK AI ...

One of its more interesting recommendations is something called Mutually Assured Compute Destruction. America would build its new datacentres in Mongolia and China would build theirs in Canada. If the agreement collapsed, each side would destroy its ...

A doctrine whose trigger mechanism can be this thoroughly disputed cannot function as a deterrent. Deterrence requires certainty. MAD worked because both sides knew, with something close to certainty, what a launch would produce. MACD, as currently s...

Since the release of the National AI Plan in December 2025, a top priority has been getting the AI Safety Institute (AISI) up and running with the right people to lead it. Because we are not building a think tank. We are building a national testing c...