Anthropic Ceo Claps Back After Trump Officials Accuse Firm Of Ai Fear-mongering  

Anthropic CEO Dario Amodei published a statement Tuesday to “set nan grounds straight” connected nan company’s alignment pinch Trump management AI policy, responding to what he called “a caller uptick successful inaccurate claims astir Anthropic’s argumentation stances.” 

“Anthropic is built connected a elemental principle: AI should beryllium a unit for human progress, not peril,” Amodei wrote. “That intends making products that are genuinely useful, speaking honestly astir risks and benefits, and moving pinch anyone superior astir getting this right.” 

Amodei’s consequence comes aft last week’s dogpiling connected Anthropic from AI leaders and apical members of nan Trump administration, including AI czar David Sacks and White House elder argumentation advisor for AI Sriram Krishnan — each accusing nan AI elephantine of stoking fears to harm nan industry.  

The first deed came from Sacks aft Anthropic co-founder Jack Clark shared his hopes and “appropriate fears” astir AI, including that AI is simply a powerful, mysterious, “somewhat unpredictable” creature, not a dependable instrumentality that’s easy mastered and put to work.  

Sacks’s response: “Anthropic is moving a blase regulatory seizure strategy based connected fear-mongering. It is principally responsible for nan authorities regulatory frenzy that is damaging nan startup ecosystem.”  

California Senator Scott Wiener, writer of AI information measure SB 53, defended Anthropic, calling retired President Trump’s “effort to prohibition states from acting connected AI w/o advancing national protections.” Sacks past doubled down, claiming Anthropic was moving pinch Wiener to “impose nan Left’s imagination of AI regulation.” 

Further commentary ensued, pinch anti-regulation advocates for illustration Groq COO Sunny Madra saying that Anthropic was “causing chaos for nan full industry” by advocating for a modicum of AI information measures alternatively of unfettered innovation. 

In his statement, Amodei said managing nan societal impacts of AI should beryllium a matter of “policy complete politics,” and that he believes everyone wants to guarantee America secures its lead successful AI improvement while besides building tech that benefits nan American people. He defended Anthropic’s alignment pinch nan Trump management successful cardinal areas of AI argumentation and called retired examples of times he personally played shot pinch nan president.  

For example, Amodei pointed to Anthropic’s activity pinch nan national government, including nan firm’s offering of Claude to nan national government and Anthropic’s $200 cardinal statement pinch nan Department of Defense (which Amodei called “the Department of War,” echoing Trump’s preferred terminology, though nan sanction alteration requires legislature approval). He besides noted that Anthropic publically praised Trump’s AI Action Plan and has been supportive of Trump’s efforts to grow power proviso to “win nan AI race.” 

Despite these shows of cooperation, Anthropic has caught power from manufacture peers from stepping extracurricular nan Silicon Valley statement connected definite argumentation issues. 

The institution first drew ire from Silicon Valley-linked officials erstwhile it opposed a proposed 10-year ban connected state-level AI regulation, a proviso that faced wide bipartisan pushback. 

Many successful Silicon Valley, including leaders astatine OpenAI, person claimed that authorities AI regularisation would slow down nan manufacture and manus China nan lead. Amodei countered that nan existent consequence is that nan U.S. continues to capable China’s information centers pinch powerful AI chips from Nvidia, adding that Anthropic restricts nan waste of its AI services to China-controlled companies contempt gross hits.  

“There are products we will not build and risks we will not take, moreover if they would make money,” Amodei said. 

Anthropic besides fell retired of favour pinch definite powerfulness players erstwhile it supported California’s SB 53, a light-touch information measure that requires nan largest AI developers to make frontier exemplary information protocols public. Amodei noted that nan measure has a carve-out for companies pinch yearly gross gross beneath $500 million, which would exempt astir startups from immoderate undue burdens.  

“Some person suggested that we are someway willing successful harming nan startup ecosystem,” Amodei wrote, referring to Sacks’s post. “Startups are among our astir important customers. We activity pinch tens of thousands of startups and partner pinch hundreds of accelerators and VCs. Claude is powering an wholly caller generation of AI-native companies. Damaging that ecosystem makes nary consciousness for us.” 

In his statement, Amodei said it has grown from a $1 cardinal to $7 cardinal run-rate complete nan past 9 months while managing to deploy “AI thoughtfully and responsibly.” 

“Anthropic is committed to constructive engagement connected matters of nationalist policy. When we agree, we opportunity so. When we don’t, we propose an replacement for consideration,” Amodei wrote. “We are going to support being honorable and straightforward, and will guidelines up for nan policies we judge are right. The stakes of this exertion are excessively awesome for america to do otherwise.”