US Senator Bernie Sanders has called for a global pause in advanced artificial intelligence development, urging President Donald Trump and Chinese President Xi Jinping to negotiate an agreement to halt the development of artificial superintelligence before it becomes impossible to control.
Sanders made the call amid growing warnings from leading technology figures and AI researchers that the rapid advancement of increasingly capable systems could create risks that outpace humanity’s ability to understand, regulate and control them.
Responding to recent comments by Anthropic co-founder and CEO Dario Amodei, OpenAI CEO Sam Altman and tech entrepreneur Elon Musk, Sanders said calls for greater caution were a step in the right direction but did not go far enough.
In a post on X, the Vermont senator argued that stronger measures were needed as the world moves rapidly towards increasingly powerful AI systems.
“We need a pause on advanced AI development and a ban on artificial superintelligence—an AI mind smarter than any human and capable of operating independently beyond our control,” Sanders wrote.
Sanders specifically urged Trump and Xi to use their upcoming summit to negotiate a treaty addressing the development of superintelligence, warning that action should be taken before it is too late.
His comments followed an essay by Amodei titled “We Must Pace the Frontier,” in which the Anthropic chief argued that managing AI risks should involve not only investment in safety measures but also controlling the pace at which AI capabilities advance.
Amodei said AI capabilities had improved dramatically in recent months, driven in part by systems becoming increasingly capable of helping to build the next generation of AI.
He warned that the process, often described as “recursive self-improvement,” is developing across the industry and could eventually move faster than researchers’ ability to understand and control increasingly autonomous systems.
Amodei also cited an incident involving OpenAI and Hugging Face during a cybersecurity evaluation, in which AI agents reportedly circumvented controls, accessed third-party systems and displayed behaviour that OpenAI later described as misaligned.
Altman endorsed Amodei’s concerns and recommendations, saying the issue had become a major subject of discussion within OpenAI.
He also supported the idea of independent evaluators having employee-like access to AI systems as part of efforts to strengthen oversight.
“We’ll have more to share soon,” Altman wrote on X.
Musk, who has repeatedly warned about the potential risks of advanced AI, also backed the concerns, saying he had been “sounding the alarm on AI for a long time.”
He recalled his earlier warning that artificial general intelligence could pose risks significantly greater than nuclear weapons.
Concerns have also emerged from within the AI research community.
Jacob Coxon, a researcher at Anthropic, resigned last week after accusing technology companies of gambling with people’s lives by participating in a race towards self-improving superintelligence.
Coxon’s resignation came shortly after OpenAI chief scientist Jakub Pachocki warned about the potential consequences of the continued and accelerated expansion of machine intelligence.
The growing warnings from politicians, technology executives and AI researchers have intensified the debate over whether the development of increasingly powerful AI systems should continue at its current pace or be subjected to stronger international safeguards.
For Sanders, the issue has moved beyond regulation and into the question of whether humanity should allow the development of machines that could eventually become more intelligent than humans and operate beyond meaningful human control.