* Originally in: SF SciFi -> Reality  
    
  Powerful AI systems can go rogue, behave in extremely dangerous ways, or  
  even  
  resist human intervention: A bill requiring AI systems to have a kill  
  switch  
  is now in Congress  
    
  Date:  
  Fri, 24 Jul 2026 13:05:00 +0000  
    
  Description:  
  The bill would give the Secretary of Homeland Security the ability to shut  
  down AI models that are advancing too quickly  
    
  FULL STORY  
  An 'unprecedented cyber incident' between OpenAI and Hugging Face  
  prompted the introduction of the bill -- A bill requiring powerful frontier  
  artificial intelligence systems to have an in-built kill switch has been  
  presented to Congress in the wake of the OpenAI cyberattack against Hugging  
  Face .  
    
  The incident saw an OpenAI model undergoing testing escaped its sandbox,  
  chained a number of zero day vulnerabilities and breached one of the  
  biggest  
  AI and machine learning companies, highlighting warnings by Anthropic and  
  OpenAI about the dangers of rapidly advancing AI technology.  The  
  bipartisan bill was introduced to congress on Thursday by Rep. Ted Lieu,  
  D-Calif., and Rep. Nathaniel Moran, R-Texas.  Unfortunately, powerful  
  AI systems can go rogue, behave in extremely dangerous ways, or even resist  
  human intervention, Lieu said in a statement. It is imperative that these  
  AI  
  systems have kill switches so we can keep this technology from causing  
  catastrophic harm, and that the federal government has the clear authority  
  and process to shut down rogue AI models.  
    
  OpenAI said that the attack was an unprecedented cyber incident, and is  
  working with Hugging Face to understand exactly how that incident took  
  place.  
    
  In June 2026, Anthropic called for a meaningful slowdown or pause on AI  
  technology development, stating that AI self-improvement was rapidly  
  increasing and that a pause would provide time to deal with its immense  
  implications.  
    
   OpenAI CEO Sam Altman warned in March that while AI can help with medical  
  and scientific breakthroughs, AI will also present new threats to society  
  that we have to address, adding that no company can handle this alone.  
    
  The kill switch bill would provide the Secretary of Homeland Security with  
  the authorization to issue a slow down or shut down of particular AI models  
  that threaten to cause catastrophic harm. The bill would also force AI  
  companies to implement cyber incident reporting and maintain forensic  
  records  
  to help the government and researchers understand similar incidents.  
    
  Stewardship means making sure humans keep the capability to control the  
  technology we build, Moran said in a statement. This is exactly the kind of  
  issue that needs serious attention and achievable policy, and Im glad to  
  work  
  across the aisle with Congressman Lieu toward a solution.  
    
  Via CNBC  
    
  Link to news story:  
    
  https://www.techradar.com/pro/security/powerful-ai-systems-can-go-rogue-  
  behave  
  -in-extremely-dangerous-ways-or-even-resist-human-intervention-a-bill-  
  requirin  
  g-ai-systems-to-have-a-kill-switch-is-now-in-congress  
    
  $$  
  --- MultiMail/DOS  
   * Origin: capitolcityonline.net * KY, USA (1:2320/107)