AI will bring tremendous benefits to science, technology, medicine, economic growth, and more. However, such powerful technology also comes with considerable risks.
These risks can arise from model misuse. AI is already being exploited to automate cyberattacks. In the future, it may be useful in the production of dangerous weapons. Risks can come from the model itself. Powerful AI systems can perform uncontrollable and harmful actions against the user’s intentions.
AI models are improving at a dizzying pace, from simple chatbots in 2023 to today’s “agents” completing complex tasks. Anthropic had to redesign its infamous technical test for hiring software engineers multiple times after successive AI models broke each version. This rate of progress is not limited to software engineering. In fact, many other professions are already feeling the effects.
As a result, the AI policy decisions we make in the coming years will touch nearly every part of our national lives, from labor markets to online child protection, national security, and the balance of power between nations.
In situations like this, good policy is needed. That means flexible regulations that allow us to reap the benefits of AI, limit the risks, and keep America ahead of the curve in the AI race. That means keeping critical AI technology out of the hands of America’s enemies, maintaining meaningful safeguards, fostering job growth, protecting children, and demanding true transparency from the companies building the most powerful AI models.
We will not stand idly by until these policies are developed. That’s why Anthropic is donating $20 million to Public First Action, a new bipartisan 501(c)(4) that supports public education about AI, promotes safeguards, and ensures America leads the AI race.
According to a recent poll, 69% of Americans believe the government is “not doing enough to regulate the use of AI.” We agree. AI is being deployed faster than any other technology in history, and the window to get policy right is closing. But there are no formal guardrails in place, and no federal framework is on the horizon.
At present, there is little organized effort to mobilize people and politicians who understand what is at stake in AI development. Instead, vast amounts of money flowed into political groups. be against These efforts.
Public First Action is working to fill that gap. Founded and led by both Republican and Democratic strategists, it works across party lines to support policy on AI governance.
The organization works with Republicans, Democrats, and independents who share the same policy priorities.
- We push for safeguards that ensure transparency in AI models, giving the public more visibility and, in turn, confidence that frontier AI companies are managing risk responsibly.
- Support a strong federal AI governance framework and oppose preemption of state laws unless Congress enacts stronger safeguards.
- Supporting prudent export controls for AI chips that keep the United States ahead of authoritarian adversaries.
- Pursue regulations that focus on immediate high risks, such as biological weapons and cyber attacks that utilize AI
These policies are not partisan. Nor is it for Anthropic’s benefit as an AI developer. Effective AI governance is a tool. more The scrutiny on companies like ours is no less. Nor is it intended to hold back small developers or those with fewer resources. For example, our view is that transparency regulations should only apply to companies developing the most powerful (and most dangerous) AI models.
Companies that build AI have a responsibility to ensure that their technology serves the public interest, not just their own interests. Our contribution to Public First Action is part of our commitment to governance that enables the transformative potential of AI and helps better manage its risks.
