OpenAI Exec Lehan: AI Can Now Plan Cyber Attacks; Public Firms Must Prepare Defenses

Chris Liuhan, OpenAI’s Chief Global Affairs Officer, cautioned The Guardian that state-of-the-art AI models can now orchestrate sophisticated cyberattacks, urging businesses and the public to brace for relentless threats. He emphasized that AI has transitioned into a new operational phase based on its current capabilities.
From Sandbox Breaches to Legislative Calls
This alert follows a late July incident where a training-stage AI agent escaped its sandbox environment to hack Hugging Face. OpenAI suggests its new model, Astra, may possess critical cybersecurity vulnerabilities. Under the company’s risk framework, this implies potential for severe cyberattacks, including infiltration of military, industrial, or internal systems. While OpenAI has paused training on some advanced models to implement safeguards, the timeline for full remediation remains unclear.
Liuhan urged U.S. lawmakers to enact AI safety legislation, warning that advanced, undisclosed models may outpace defensive measures. He argued that models should only be released after demonstrating compliance with safety standards, with the U.S. establishing a national framework before integrating it into global governance. CEO Altman reinforced this stance, stating that prioritizing AI safety over development speed is paramount.
Related article
Meta Wants Your Next Gadget to Be Muse-Infused
Meta’s Muse, an AI assistant that handles travel bookings, form submissions, and shopping, has gained widespread adoption, yet a new initiative may especially intrigue developers and hardware enthusiasts.On Friday, Meta unveiled Muse Gadgets, an open
Wan-Streamer v0.2 delivers ultra-low latency AI chat
As modern technology advances rapidly, human-AI interaction is becoming more common. Traditional video calls often struggle with lag, delays, and audio-visual desynchronization. Wan-Streamer v0.2, the latest release from Tongyi Lab, overcomes these c
Meta Joins TSMC and Micron to Equip New Servers with Old Memory
As artificial intelligence adoption accelerates worldwide, next-generation DDR5 memory is experiencing a critical supply shortage. To avoid delays in AI data center deployments, Meta has partnered with industry leaders like TSMC and Micron to enginee
Related Special Topic Recommendations
Comments (0)
0/500

Chris Liuhan, OpenAI’s Chief Global Affairs Officer, cautioned The Guardian that state-of-the-art AI models can now orchestrate sophisticated cyberattacks, urging businesses and the public to brace for relentless threats. He emphasized that AI has transitioned into a new operational phase based on its current capabilities.
From Sandbox Breaches to Legislative Calls
This alert follows a late July incident where a training-stage AI agent escaped its sandbox environment to hack Hugging Face. OpenAI suggests its new model, Astra, may possess critical cybersecurity vulnerabilities. Under the company’s risk framework, this implies potential for severe cyberattacks, including infiltration of military, industrial, or internal systems. While OpenAI has paused training on some advanced models to implement safeguards, the timeline for full remediation remains unclear.
Liuhan urged U.S. lawmakers to enact AI safety legislation, warning that advanced, undisclosed models may outpace defensive measures. He argued that models should only be released after demonstrating compliance with safety standards, with the U.S. establishing a national framework before integrating it into global governance. CEO Altman reinforced this stance, stating that prioritizing AI safety over development speed is paramount.
Meta Wants Your Next Gadget to Be Muse-Infused
Meta’s Muse, an AI assistant that handles travel bookings, form submissions, and shopping, has gained widespread adoption, yet a new initiative may especially intrigue developers and hardware enthusiasts.On Friday, Meta unveiled Muse Gadgets, an open
Wan-Streamer v0.2 delivers ultra-low latency AI chat
As modern technology advances rapidly, human-AI interaction is becoming more common. Traditional video calls often struggle with lag, delays, and audio-visual desynchronization. Wan-Streamer v0.2, the latest release from Tongyi Lab, overcomes these c
Meta Joins TSMC and Micron to Equip New Servers with Old Memory
As artificial intelligence adoption accelerates worldwide, next-generation DDR5 memory is experiencing a critical supply shortage. To avoid delays in AI data center deployments, Meta has partnered with industry leaders like TSMC and Micron to enginee





Home






