Humans First proposes AI export controls and H-1B visa halt

Humans First’s proposals span AI development, deployment and accountability. The framework calls for independent testing of large AI companies, while the group also proposes mandatory testing of AI models for extreme risks.[1][2] It wants mandatory public reporting of AI incidents and proposes holding frontier AI developers liable for reckless conduct.[1] The group also calls for strict rules governing AI used by children.[1] Together, these proposals would put testing and disclosure on the agenda alongside potential legal responsibility; the evidence describes them as proposals, not enacted requirements.[1][2]

The platform takes positions on both AI capabilities and government technology use. Humans First wants to prohibit advanced AI practices such as recursive self-improvement, and The Hill describes the group as calling for a “red line” on self-improving AI and limits on developing “vastly superhuman AI.”[1][2] The group also wants to ban federal use of cameras like Flock.[1] Separately, it calls for a pause on H-1B visas; The Hill characterizes the proposal as a halt to approvals.[1][2] The Hill’s report also mentions warrantless searches of private data in a quote.[2]

Humans First initially organized across party lines, then shifted its focus to conservative activists in April, when it named Amy Kremer chair.[2] Kremer is a longtime Tea Party activist, according to Semafor.[1] The group led a day of nationwide demonstrations in July and staged a cross-country bus tour opposing AI infrastructure that ended earlier in October.[2] Its platform says policy should center on the American people rather than globalist corporations or unaccountable bureaucrats, and urges Congress to act and conservatives to lead.[1]

Semafor reports that the framework goes beyond proposals from some Republicans interested in regulating AI and further than the Trump administration’s light-touch approach.[1] The Hill reports that Trump argued stricter regulations could hamper the United States’ ability to compete with China.[2] AI leaders have signed on to voluntary commitments for internal model controls and external auditing, The Hill reports.[2] The evidence does not say that Humans First’s proposals have been adopted or that they represent the position of the administration or AI companies.[1][2]