Anthropic’s push for AI regulation has become a lightning rod in Washington, drawing comparisons to the hardball influence tactics once used by Sam Bankman-Fried. The company says strict guardrails are needed to keep powerful models from causing real harm, while critics see a familiar pattern of big spending, safety talk, and policy pressure that could tilt the field in its favor.
Anthropic, the company behind the Claude chatbot, has leaned hard into the idea that artificial intelligence needs tighter federal oversight. Dario Amodei has repeatedly warned that advanced systems could create national security and public safety risks, and he has argued for mandatory testing and auditing before the most powerful models are deployed.
That message has landed in a tense political moment. The Trump administration has made AI leadership a priority, and many in Washington are eager to avoid falling behind China and other rivals, which gives safety arguments extra weight and a lot more room to shape the debate.
Still, the comparison to Bankman-Fried keeps popping up because the script looks oddly familiar. Before his collapse, the former FTX chief pushed for crypto rules under the banner of consumer protection, all while spending heavily on political access and building a network that critics said was designed to help his own business.
Bankman-Fried later claimed his public support for regulation was “just PR” in a direct-message exchange that surfaced after FTX imploded. He had backed legislation that would have given the Commodity Futures Trading Commission a central role in overseeing crypto, and prosecutors later said his political operation was tied to illegal contributions and an effort to expand his influence in Washington.
Anthropic insists that its approach is different, and there are obvious legal distinctions between disclosed lobbying and a criminal scheme. But the core dynamic still raises eyebrows, because the company has spent millions on lobbying, supported pro-regulation groups, and backed a broader push for tighter rules that could end up shaping the market in its favor.
Critics argue that this is where the story gets sticky. When a well-funded company warns that the industry is too dangerous to leave alone, the policy debate can start to feel less like public service and more like a competitive weapon.
Those concerns have sharpened around California’s SB 53, a proposal that some say could hit smaller AI labs harder than the giants. Amodei has pushed back by saying startups matter too and that the company’s safety message is not meant to crush rivals, but skeptics see a regulatory framework that could raise costs and make life easier for the biggest names in the room.
That tension is why Anthropic’s spending keeps getting attention. Federal disclosures show the company has invested heavily in lobbying, and it has also set aside major funding for outside groups that can amplify pro-regulation arguments in ways that sound a lot like a coordinated campaign.
David Sacks, a former Trump administration AI adviser, blasted the company’s approach as “a sophisticated regulatory capture strategy based on fear-mongering” and argued that the panic around AI is hurting the startup ecosystem. That kind of criticism reflects a broader suspicion in tech circles that the loudest calls for safety are not always coming from the most selfless place.
Amodei has tried to keep the focus on the risks he believes are real, saying AI models may need tough oversight before they become even more powerful. He told the Senate Judiciary Committee that AI poses “extraordinarily grave threats to U.S. national security” and called for a mandatory “testing and auditing regime” with the power to stop unsafe systems from going live.
That sort of language carries weight because it taps into fears that feel bigger than a standard policy fight. If AI can actually go rogue, then lawmakers are dealing with something that looks less like a normal tech dispute and more like a race to stop the next major disaster before it starts.
At the same time, the industry has not forgotten how often regulation gets used as a shield by companies that want to freeze the competition. The comparison to FTX stings because it suggests that some of the loudest guardians of public safety may also be the ones best positioned to benefit when the rules get tougher.
Anthropic’s defenders say that framing is too cynical and misses the genuine risks tied to frontier AI. Its critics say the pattern is too neat to ignore, especially when the same company warning about dangers is also spending heavily to shape the laws meant to address them.
That clash is not going away anytime soon, because AI policy is now tied to national power, business strategy, and public fear all at once. In Washington, that mix usually means one thing: the fight over who gets to write the rules is just getting started.
