Salesforce staff press Benioff to bar ICE from using company technology after event remarks

A group of Salesforce employees circulated an internal appeal pressing CEO Marc Benioff to refuse to enable U.S. Immigration and Customs Enforcement (ICE) with Salesforce cloud and AI products. The campaign gained traction after Benioff’s remarks at a company leadership event provoked immediate internal backlash and revived sensitivities about the company’s political interventions. Staffers point to leaked documentation describing a proposed initiative that would help ICE recruit roughly 10,000 agents and process tip-line information at scale, framing that plan as a material shift in how Salesforce technology could be applied to immigration enforcement. The petition asks Benioff to use his influence in Washington to condemn ICE’s practices publicly and to establish a clear corporate policy barring the deployment of Salesforce systems for enforcement operations. The memo also highlights reporting on detention practices, referencing approximately 66,000 detainees and the claim that a large share lack criminal convictions. Organizers say the request is both ethical and operational: they argue enabling tools that scale enforcement would contradict Salesforce’s stated commitments on responsible AI and human-centered technology. Employees warned they are prepared to escalate pressure through further organizing and public statements if leadership declines to adopt restrictive terms. The push reflects a broader trend inside tech firms, where workers have increasingly demanded transparency about government contracts and sought stricter limits on law-enforcement and immigration-related uses of advanced systems. Executives at other AI companies have offered measured public rebukes in past incidents, balancing workforce expectations with customer relationships and access to policymakers; organizers say such calibrated responses often leave employees unsatisfied. The dispute raises immediate reputational risk for Salesforce and could complicate government partnerships if leadership resists the demands. Legal and political channels — including state-level litigation, federal inquiries, and contract-disclosure pressure — give activists additional leverage beyond internal protest. The episode underscores the evolving governance challenge for enterprise AI vendors: companies are being asked to more precisely define permissible government uses and to codify those boundaries in product and contracting terms.