Anthropic

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.

Public input is critical to ensuring that powerful AI serves humanity’s interests.

Our Advanced AI Framework proposes mandatory independent safety testing for frontier models, transparency requirements, and government authority to block or recall dangerous AI deployments.

Frontier developers should engage at least one qualified independent evaluator to publish a review of their evaluations and risk reports. In parallel, governments and industry should develop the ecosystem of qualified independent evaluators by settin...

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.

Our Advanced AI Framework proposes mandatory independent safety testing for frontier models, transparency requirements, and government authority to block or recall dangerous AI deployments.

We believe it would be good for the world to have the option to slow or temporarily pause frontier AI development to enable societal structures and alignment research to keep up with the advance of the technology.

If AI delivers even a fraction of its potential, it could generate unprecedented abundance. We could see decades of scientific progress compressed into years, and levels of productivity growth that today seem implausible.

Equity funding: policymakers should expand the mechanisms by which these accounts can be funded, including with equity in AI companies, so that beneficiaries share directly in nearer-term gains from AI-driven growth.

In the “unprecedented” scenario, the company wrote that more permanent support will be necessary, and it listed several ways to generate and share revenue broadly, including basic income, sovereign wealth models and equity-sharing mechanisms.

Because effective tax rates on labor substantially exceed those on capital, a shift in national income from labor toward capital means the current tax system captures a shrinking share of a growing economy, even as the cost of sustained income suppor...

Developers should test their models against each of the four Enumerated Risk categories and publish a summary of the results.

Require Covered Developers to report Critical Safety Incidents to the designated Agency within 15 days of the Covered Developer discovering the Critical Safety Incident or facts that would lead the Covered Developer to reasonably conclude that the Cr...

Establish universal pre-distributive capital accounts. We should give every American a direct financial stake in the economy at the moment when AI is fueling growth.

Absent a strong federal law, states should retain full authority to legislate.

Champion the export of American AI. As public and commercial sectors around the world increasingly adopt AI, the Trump administration should continue its efforts to promote the global adoption of trusted AI hardware and models developed and shaped by...

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...

If policymakers in the US and allied democracies act to close these two channels propping up China’s AI models—illicit and evasive compute access and illicit model access—then we have a potentially once-in-a-generation opportunity to secure our lead.

It’s built on five core principles: keeping humans in control, aligning with human values, securing agents’ interactions, maintaining transparency, and protecting privacy.

We find no systematic increase in unemployment for highly exposed workers since late 2022, though we find suggestive evidence that hiring of younger workers has slowed in exposed occupations.

Second, we believe that mass domestic surveillance of Americans constitutes a violation of fundamental rights.

We do not believe that today’s frontier AI models are reliable enough to be used in fully autonomous weapons. Allowing current models to be used in this way would endanger America’s warfighters and civilians.

Mass domestic surveillance. We support the use of AI for lawful foreign intelligence and counterintelligence missions. But using these systems for mass domestic surveillance is incompatible with democratic values.

We are not sure whether Claude is a moral patient, and if it is, what kind of weight its interests warrant. But we think the issue is live enough to warrant caution, which is reflected in our ongoing efforts on model welfare.

Before approving a new model for easy or broad internal use (e.g., making the model available via our most-used internal tools), we conduct a pre-deployment review process.

We treat misalignment as a central, but solvable, problem for both us and for the field.

But if that happens, it's likely that AI will have acted as a general substitute for labor: these gains could come with a reduction in demand for human work that is difficult or impossible for many workers to absorb. In this world, the central econom...

If it were possible to effectively slow the development of this technology to give ourselves more time to deal with its immense implications, we think that would likely be a good thing. But if a slowdown simply lets the least cautious actors catch up...

Importantly, the law balances [...] whistleblower protections—[...]. [...] Protecting whistleblowers: It should be an explicit violation of law [...] punish employees who raise concerns [...].

Our proposed transparency framework offers a practical first step: public visibility into safety practices while preserving private sector agility to deliver AI's transformative potential.

If you decide to turn off the model training setting, we will not use any new chats [...] for future model training.

ASL-2 refers to systems that show early signs of dangerous capabilities – [...] give instructions on how to build bioweapons – [...] e.g. a search engine couldn’t.

we propose that AI developers pre-register large training runs [...] within their home country’s government [...] plausible bar for now would be 10^26 FLOP or higher.