Trump calls AI safety a hoax as industry urges safeguards
- US President Donald Trump characterised AI safety concerns as a "hoax" in social media posts on 15 September 2026, comparing warnings to climate change scepticism.
- Anthropic, OpenAI and xAI have publicly called for development slowdowns and closer monitoring, with Anthropic's Jack Clark arguing that a verifiable "kill switch" may require legislation.
- An Anthropic scientist's estimate of a greater than ten per cent probability of AI-caused human extinction within a decade contributed to a technology sector share sell-off.
- Former Trump strategist Steve Bannon and Senator Bernie Sanders were scheduled to address a bipartisan "Pro-Human Assembly" on AI governance in Washington.
Donald Trump dismissed warnings about artificial intelligence safety as a “hoax” in social media posts on 15 September 2026, comparing concerns about the technology to climate change and the Russia investigation. The remarks landed at a moment when the technology industry’s own leaders were publicly urging slower development and stronger oversight, an unusual split between the executive branch and the sector it regulates.
The strongest version of Trump’s position holds that anxiety about AI is politically motivated and economically self-serving. Industry figures who warn about existential risk are, in this framing, either hyping their own importance or trying to erect barriers that would advantage incumbents. Regulation becomes a constraint on American competitiveness, particularly vis-à-vis China, where state media characterised the US debate as driven by commercial rivalry rather than genuine safety concerns.
The opposing case comes from the companies actually building these systems. Jack Clark, co-founder of Anthropic, told the BBC’s report on Trump says AI safety fears a ‘hoax’ as he rejects calls for greater safeguards that a third-party-verifiable “kill switch” for dangerous AI models may eventually require legislation, noting that most laboratories already maintain internal shutdown mechanisms. Anthropic’s chief executive Dario Amodei has called for a development slowdown and closer monitoring, a position echoed by OpenAI’s Sam Altman and Elon Musk’s xAI. Microsoft AI released guidance on what it termed “humanist AI”, with chief executive Mustafa Suleyman saying the document had been prepared over several months. On the political front, former Trump strategist Steve Bannon and Senator Bernie Sanders were scheduled to address a bipartisan “Pro-Human Assembly” in Washington, an alignment that underscores how far safety concern has spread beyond its traditional academic base.
The convergence is real, if uncomfortable. When the firms most financially exposed to deregulation are themselves requesting guardrails, the argument that safety fears are a competitive hoax becomes harder to sustain without engaging the substance. An Anthropic scientist’s estimate that AI-caused human extinction carries a probability exceeding ten per cent within a decade contributed to a share sell-off in some technology firms on 15 September, a reminder that these are not abstract policy debates but claims with immediate market consequences. The bipartisan event in Washington, pairing Bannon and Sanders, suggests AI governance is settling into a category of issue where left-right agreement is possible, even if the executive branch resists.
The unresolved question is whether a deregulatory posture can coexist with an industry that is, for the first time, asking to be constrained. If the companies building frontier models are correct that meaningful safety requires external verification and legislative backing, dismissing their concerns as a hoax does not eliminate the risk. It merely removes the possibility of managing it collectively.