Expert opinions on AI Safety

Every AI agent should have its own identity and access to only what it needs. [...] High-impact or irreversible actions should require human approval.

Existing responsible AI governance frameworks mandate human oversight. Yet they share a critical assumption: that competent human overseers remain present and capable of exercising meaningful oversight and control over AI decisions. Accountability mu...

A lot of tools can tell you an AI agent is behaving strangely. Very few can act on that at the network layer, without waiting on a human to approve a ticket. That’s the piece we’re closing.

We have slowed down reinforcement learning and training runs because we do believe that safety should be of utmost priority. Of course, we want to hit AGI as soon as possible and we want to get there first. But we also want to make sure that AGI bene...

Governments must assume their responsibilities. It is the first responsibility of any government to protect its citizens – including from the threats of artificial intelligence. National action is essential. But global coordination is also indispensa...

We need to think about the regulation. I don't think we need to stifle it or stop the growth of it, but equally, it will be foolish for anybody to think that we can let it flow to a level, and then if an accident happens, then we'll think about it.

States have a responsibility – to their own populations, and to those in other states who will face the consequences of the proliferation and use of AI-enabled weapons and by conflict escalation – to set clear limits. A company cannot determine what ...

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 United States totally rejects any attempt to construct a globalist scheme of control for the artificial intelligence being spoken of so much now – hereinafter officially called super intelligence.

We need to be really careful with open-source on safety. I can't talk for open models. But by definition, there's more risk with them.

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...

If safeguards are properly followed, businesses, including banks, should have enough guarantees to essentially run any model, regardless of where they come from.

Embedded evaluators are independent external experts placed within frontier AI companies to provide ongoing assessments and validation of a company’s safety, security, and operational systems, practices, and claims. [...] Contracts should contain spe...

Connecticut already has strong whistleblower laws that protect employees from retaliation. [...] creating a separate, AI-specific whistleblower regime could lead to confusion, duplication, and unintended consequences for employers operating in highly...

Embedded evaluators are independent external experts placed within frontier AI companies to provide ongoing assessments and validation of a company’s safety, security, and operational systems, practices, and claims. [...] As frontier AI companies beg...

the bad news: Dario estimates that in 6 months AI will have the ability to take over the internet the worse news: Anthropic's current commitment to stopping this is... putting independent evaluators in desks in their office have a great day

They really become part of the company, which makes it difficult to criticize it because the staffers become almost your friends.

Society may eventually find it useful to create a category of limited artificial legal personality without deciding that the entities enjoying it possess subjective experience. [...] Corporations already demonstrate that legal persons can enjoy only ...

There is a reasonable chance, however, that if the vision for AI agents comes to pass, then for our markets and legal accountability regimes to function effectively, agents will have to have the capacity to sue and be sued directly. This is the defin...

Either way, an AI agent is an automated tool, not a legal entity, so accountability always lands on the human administrators who configured its permissions.

Internal use: The same systemic-risk assessment applies to internal use of frontier models, including research, operational systems, and internal products. Internal deployment is assessed against the same capability thresholds and safeguards as pre-r...

Such a deal could start with a government-to-government framework that affords Australia early and preferential access to AI frontier models and a guarantee of continued access.

Allies should establish secure mechanisms to share information on serious AI incidents, emerging catastrophic capabilities, evaluation methods and effective safeguards.

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...

The first is a federal governance framework targeting frontier AI developers above $500 million in annual revenue. These developers would be required to publish safety frameworks, report critical incidents, and submit to semi-annual third-party audit...

The FRONTIER Act is a targeted bill with clear, uniform transparency, and safety standards, instead of a fifty-state patchwork. I'm glad to support this bipartisan effort so that Congress acts before this technology outruns us.

Physicists faced a version of this problem in the 1950s. No single lab could afford the machine they needed, so they pooled resources and built one together, with shared data and shared credit. This piece sketches what the same move could look like f...

It is irresponsible for society to allow them to move forward and make these products even more advanced. That’s why I am introducing legislation to immediately pause the development of increasingly powerful AI and ban the creation of systems that hu...

A frontier developer shall report, using the mechanism established under subsection (a), any critical safety incident pertaining to one or more of its frontier models to the agency and the Attorney General within 72 hours of the frontier developer le...

Timelines for incident reports are often set with well-intentioned urgency, but if they implicitly assume that root-cause analysis or full impact assessment has already been conducted, they will either produce premature and unreliable reports or plac...

To maximize the effectiveness of such provisions, policymakers should define a sensible range of time frames to report incidents, possibly with a graduated approach according to the gravity of the harm.

For a long time many of us thought of RSI as the way AGI turns into superintelligence. [...] I was advocating what I called a semi-hard takeoff: a period where the self-improvement was real but not that fast, followed by a point -- once the AGI fully...

I do not think this will happen. I don’t think that automating AI R&D will rapidly produce domain-general superintelligence. I don’t think that a data centre of geniuses, largely separated from the rest of the economy, will figure out how to cure Alz...

AI is already on the verge of going rogue, but with legal person status, it could become an actual existential threat to humanity.

We have to change the way we talk about AI risk because for a very long time, it’s either existential and everything is going to die and the killer robots are coming, or it doesn’t really matter.

We do need to have some safeguards in place to make sure there are no unintended consequences, but, no, I do not think AI is going to imperil human existence.