The leaders of Anthropic and OpenAI sharply criticized recent violence by federal immigration authorities while simultaneously offering supportive words for President Trump, a split-screen message that has stirred unease across the tech industry and among their own employees.
Public Rebukes Follow Minneapolis Killings
On NBC News, Anthropic CEO Dario Amodei warned that images of federal agents using deadly force in Minneapolis underscored the need to defend democratic values at home, not just abroad. He emphasized that Anthropic holds no contracts with Immigration and Customs Enforcement and urged accountability as videos of one victim’s killing spread widely online.
At OpenAI, a leaked Slack message reported by The New York Times showed CEO Sam Altman telling employees that ICE actions had crossed a line. He contrasted the removal of violent offenders with sweeping enforcement tactics in cities, framing it as a matter of American civic responsibility to push back against overreach.
Their statements track a familiar pattern in Silicon Valley, where worker activism often forces executives to clarify stances on government work. In recent years, rank-and-file employees at major tech firms have protested contracts linked to immigration enforcement and military uses of AI, recalling episodes from Google’s Project Maven to corporate boycotts tied to family separation policies.
Praise for Trump Draws Backlash Across Tech Industry
What made these remarks notable was the accompanying praise for Trump. Amodei welcomed the president’s consideration of allowing Minnesota authorities to run an independent probe into the shootings by federal agents, a move that some Republicans have begun to support as pressure for transparency grows.
Altman, in his internal note, expressed hope that Trump, described as a strong leader, would unify the country in response to the crisis. That balancing act drew criticism from some industry voices. J.J. Colao of Haymaker Group, a signatory to an open letter at ICEout.tech, said the rhetoric risked letting the White House evade responsibility even as the statement helped spotlight abuses.
The shift is particularly striking given Altman’s past writing that was searingly critical of Trump ahead of his first term. Amodei, for his part, recently blasted proposals to let advanced U.S. AI chips flow to China, likening the idea to arming an adversary—evidence that neither executive shies from high-stakes policy fights.
Industry Pressure and Contract Questions
The employee push is direct: workers have urged their CEOs to press the White House to pull federal agents from U.S. cities and to cancel any contracts with ICE. Amodei’s declaration that Anthropic has none answers part of that call; OpenAI has not publicly detailed contract exposure but told staff it would engage policymakers and “do the right thing.”
Civil rights groups have documented long-running concerns with immigration enforcement tactics. ICE’s Enforcement and Removal Operations reported roughly 170,000 administrative arrests and about 140,000 removals in FY2023, according to DHS figures, with watchdogs arguing that local deployments and coordination can expand beyond the narrow focus on dangerous offenders. The Minneapolis killings, captured on video, have intensified scrutiny of such operations inside U.S. cities.
For tech firms, the risk is twofold: reputational damage with talent and users, and potential entanglement in government programs that run counter to stated AI safety commitments. Most leading labs maintain “constitution” or “safety” frameworks for deployment; critics say those commitments ring hollow if companies are silent about how their tools may be used by agencies facing misconduct allegations.
AI Boom and Political Calculus Shape CEO Responses
The context is a torrid AI funding cycle and an administration that has signaled openness to accelerated AI deployment. Industry reports have pegged OpenAI’s fundraising ambitions in the tens of billions with a sky-high private valuation, while Anthropic has reportedly raised multibillion-dollar rounds and pursued additional capital at steep valuations. Favorable policy signals—from export rules to federal AI adoption—have been a tailwind.
That backdrop helps explain why CEOs may be threading a needle: condemning specific abuses by ICE and Border Patrol while keeping lines open with a White House whose decisions shape compute access, guardrails, and procurement. Investors, employees, and regulators are all watching how AI leaders navigate that tension.
What to Watch Next in the Tech and Policy Fallout
Key questions now hinge on follow-through.
- Will the administration authorize a truly independent investigation in Minnesota?
- Will OpenAI and Anthropic commit to clearer rules governing any work with immigration or law enforcement agencies?
- Will other AI leaders join calls for limits on domestic deployments of federal agents?
For a workforce newly energized by AI’s societal stakes, statements are table stakes. Employees want action—policy engagement, contract transparency, and, where necessary, refusal. As the nation debates the bounds of immigration enforcement, the tech sector’s most powerful voices are being pressed to match their words with concrete steps.
