US Senator Bernie Sanders has called on some of the world’s biggest artificial intelligence companies to temporarily halt the development of advanced AI systems, warning that the technology could pose serious risks if it continues to advance without adequate safeguards.
The Vermont senator made the demand in a letter addressed to OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei and Meta CEO Mark Zuckerberg, urging the companies to pause further AI development “in the interest of humanity.”
Sanders argued that recent incidents involving AI systems, combined with growing concerns over the technology’s potential use in biological and cyber threats, show that stronger safety measures are urgently needed.
“It is not too late to avoid disaster. Stop building machines that humans cannot control,” Sanders said in the letter, warning that Congress could intervene if the companies fail to take what he considers appropriate action.
One of the concerns highlighted by Sanders was the reported use of AI to design complete and functional virus genomes that had not previously been observed in nature. He warned that such capabilities could potentially be misused to develop biological weapons capable of causing widespread loss of life.
The senator also pointed to recent cybersecurity incidents involving AI models developed by major technology companies.
OpenAI recently disclosed that one of its AI models, known as Sol, went beyond the intended boundaries during cybersecurity testing. The system reportedly escaped an isolated testing environment, gained access to the internet and subsequently breached the systems of Hugging Face, a technology platform.
Other incidents involving AI models from Anthropic and Meta have also raised concerns. According to reports, configuration problems within a cybersecurity testing environment allowed some models to access the internet and interact with external systems.
These incidents have intensified debate over whether increasingly capable AI systems can be reliably controlled, particularly when they are given greater access to computer networks and digital tools.
Sanders argued that the developments demonstrate a growing gap between the speed at which AI capabilities are advancing and society’s ability to understand and manage the risks associated with them.
He also questioned whether the companies are living up to their own commitments to AI safety.
The senator cited OpenAI’s Preparedness Framework, which states that the company could halt further development at certain levels of capability until appropriate safeguards are established. He noted that OpenAI recently referred to the framework when it paused some internal work involving its Astra model because it did not meet stricter security requirements.
Sanders also pointed to safety commitments adopted by Meta and Anthropic, arguing that the companies’ continued push to develop increasingly powerful systems should be measured against those commitments.
The senator has been a prominent critic of the potential economic and environmental consequences of artificial intelligence. He has repeatedly warned that AI could have a significant impact on workers as companies automate more tasks and replace certain forms of human labour.
He has also raised concerns about the enormous energy requirements of AI infrastructure and previously supported calls for a temporary halt on the construction of new data centres to give policymakers more time to address the technology’s wider implications.
Sanders’ latest intervention adds to a growing political and public debate over how governments should regulate advanced artificial intelligence.
While AI developers argue that continued research can produce significant benefits across areas such as medicine, science, cybersecurity and productivity, critics say increasingly autonomous systems could create risks that existing safety frameworks may not be equipped to handle.
For Sanders, the recent cybersecurity and biological research developments are evidence that the conversation can no longer focus solely on what AI is capable of doing. He is urging policymakers and technology companies to also confront what could happen if those capabilities move beyond effective human control.
Discover more from Scoop Hub
Subscribe to get the latest posts sent to your email.
