Expert opinions on Policy

Advanced model weights can be high-consequence assets. If stolen or openly released without adequate safeguards, they may allow actors to remove safety filters, fine-tune models for harmful uses, or reproduce capabilities that would otherwise remain ...

A safety regime that restricts legitimate defenders, while capable models remain available for attackers, creates an asymmetric disadvantage.

Tax policy can be an important tool to generate tax revenue and address growing inequality in the age of AI. Economists have studied how tax policy can actively shape labor markets. Daron Acemoglu, Andrea Manera, and Pascual Restrepo found the U.S. t...

Raising the cost of all those innovation- and productivity-enhancing uses of computing would unequivocally harm American productivity, innovation, and global competitiveness. Knowingly imposing these costs on ourselves because of concerns about certa...

Pay closer attention to the systemic risk posed by unlabelled AI-generated content, introduce stricter labelling obligations on technology services producing, deploying, or distributing AI-generated content, as well as formulate binding requirements ...

We should ask ourselves an important question: does the problem require government regulation?

Without a shared governance baseline, AI in our sector might develop in proprietary and uneven ways, deepening the very divides the panel describes and leaving smaller and developing country operators behind.

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...

Because we have the same rules applicable everywhere in the world. No matter if you take a plane in Kabul or if you take a plane in New York, the same safety rules apply everywhere. This is exactly what we need to make AI a success for business and f...

A one-size-fits-all approach to AI governance will not be effective. We all face different risks, priorities, levels of digital development, and regulatory environments. Effective governance requires shared principles that can be implemented flexibly...

Promising directions we’ve identified so far include developing the capacity to detect and respond to AI systems acting outside their developers' control, and building infrastructure for containing or shutting down such systems.

If you’re relying on a ‘kill switch’ to stop an AI agent, you’ve already made a fundamental mistake. A properly governed agent is simply never handed the task, so there is nothing to kill.

Large language models simply shouldn’t be used for targeting, since they will inevitably hallucinate and contribute to civilian harm.

The Government strongly believes that there is a need to maintain meaningful human control of all weapons systems, including any systems that may be developed. Our position is that, for such systems to comply with international humanitarian law, cont...

I need a reliable, steady partner that gives me something, that'll work with me on autonomous, because someday it'll be real and we're starting to see earlier versions of that.

Congress needs to listen to the American people and act urgently to stop the unholy alliance of AI and data brokers to fuel war and mass surveillance.

This has effectively given police a very long leash to experiment with these technologies. And now we're in a tipping point situation where it's about to be expanded all over the country, and there's a real necessity for those legal frameworks to be ...

When I was in justice, my ultimate vision for that part of the criminal justice system was to achieve, by means of AI and technology, what Jeremy Bentham tried to do with his Panopticon. That is that the eyes of the state can be on you at all times. ...

State regulation remains the only meaningful safeguard standing between Americans and a rapidly expanding set of AI-related harms.

[The bill] would create a standalone state framework that increases compliance burdens without delivering clear safety benefits.

Even after 3 years working on AI capability evaluations, the recent pace of progress has been a shock. To avoid catastrophic harms from AGI, we need much, much better coordination.

Much has been said about whether AI poses an ultimate existential threat to humanity. The more immediate existential challenge may be whether technologies that simulate relationships weaken our capacity for the relationships that foster our own human...

By framing AI as a mysterious entity, an existential threat, a utopian savior, or a moral scapegoat, these narratives obscure the reality of AI as a statistical tool.

Governor Pritzker does not believe big tech companies should ever be given a full shield that evades responsibilities they should have to protect the public interest.

Existing legal principles, such as under tort and contract law, can be applied by our courts to assess and determine liability in appropriate cases. That being said, we share the concerns raised by the Member, and will continue to study this issue, i...

States should not be permitted to penalize AI developers for a third party’s unlawful conduct involving their models.

Compounding the problem, labs in China often release dual-use capable models as open-weight. Once a model is open-weight, safeguards that do exist can be removed, making the model available to any state or non-state actor to use for malicious purpose...

Restricting access to a dual-use AI model is precautionary only if it delays harmful actors more than defenders. That condition varies across actors: a state agency or organized criminal group may obtain a substitute through theft, distillation, inte...

Shared access to open-source AI systems is important for protection in a world where AI can be used to hack into electronic systems, like voting systems and benefits infrastructure.

Dedicated AIRR capacity supports the AI Security Institute and the Sovereign AI Unit, including guaranteed resource for mission-critical AI safety research, and strategically selected UK AI startups.

We need to slow down the development of AI around the world, and we need to bring the international community together to address the risk posed by AI.

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 conclusion is sobering: a global stop to the development of superintelligent AI seems impossible.

We prohibit the use of our technology for mass surveillance of civilian populations, require compliance with law and contract, and use internal review mechanisms to assess and address higher-risk scenarios.

The use of live facial recognition in our public spaces is an extraordinary expansion of state power that currently exists in a legal vacuum. We are not Luddites on these Benches; we recognise the utility of technology, but we must ensure that live f...

The new legal framework must include a meaningful regulatory regime which operationalises principles such as necessity, responsibility, proportionality, transparency, accountability and lawfulness. These principles should then structure and constrain...