Anthropic CEO Dario Amodei issued a statement Tuesday to address what he described as "a recent uptick in inaccurate claims about Anthropic’s policy stances," particularly concerning the company's alignment with U.S. AI policy.
Amodei's statement followed a period of public criticism from figures associated with the Trump administration and other AI industry leaders. David Sacks, identified as the Trump administration's AI czar, and Sriram Krishnan, White House senior policy advisor for AI, accused Anthropic of engaging in "fear-mongering" and a "sophisticated regulatory capture strategy" that reportedly damages the startup ecosystem. These accusations emerged after Anthropic co-founder Jack Clark articulated concerns regarding AI's powerful and unpredictable nature.
In his response, Amodei asserted that Anthropic operates on the principle that "AI should be a force for human progress, not peril." He defended the company's approach to AI safety, stating that managing the societal impacts of AI should prioritize "policy over politics." Amodei highlighted Anthropic's cooperation with the federal government, including offering its Claude AI model to all three branches of government and a $200 million agreement with the Department of Defense. He also noted Anthropic's public praise for Trump’s AI Action Plan and support for efforts to expand energy provisions to advance AI development.
Despite these collaborations, Anthropic has diverged from some industry peers on specific policy issues. The company opposed a proposed 10-year ban on state-level AI regulation, a stance that drew pushback from elements within Silicon Valley, including some at OpenAI, who contend that state regulations could impede industry progress. Amodei countered that a greater risk involves the continued supply of powerful AI chips to China and affirmed Anthropic's policy of restricting sales of its AI services to China-controlled entities, even at the cost of revenue.
Anthropic also supported California's SB 53, a safety bill requiring large AI developers to make frontier model safety protocols public, which includes a carve-out for startups below a $500 million annual gross revenue threshold. Amodei refuted claims that Anthropic seeks to harm the startup ecosystem, emphasizing that startups are key customers for the company's Claude AI.
Amodei reported that Anthropic's run-rate has grown from $1 billion to $7 billion over the past nine months, coinciding with its efforts to deploy AI "thoughtfully and responsibly." He concluded by reiterating Anthropic's commitment to constructive engagement on public policy, stating the company will continue to advocate for what it believes are appropriate policies given the technology's implications.
