MI5 Launches AI Threat Centre Amid Terrorism Risks

Key Takeaways

• MI5 chief Ken McCallum warns AI is supercharging terrorism and bioweapon risks in the UK
• AI Threat Assessment Centre (AITAC) launched to assess AI-driven national security threats
• Terrorists caught using AI-enabled drones and AI tools to improve attacks
• UK names Russia as biggest threat, with state threat cases rising fourfold in a decade

MI5 chief warns AI is supercharging national security threats

MI5 Director General Ken McCallum warned that artificial intelligence is supercharging cybersecurity threats in the UK. He said terrorists have been caught using AI to improve their reconnaissance and attack methods, including with AI-enabled drones. AI could also change the stakes on bioweapons. McCallum announced the creation of the AI Threat Assessment Centre, or AITAC, to better assess these risks. He also named Russia as the biggest threat to the UK.

UK spy chief warns AI could change bioweapon risks

MI5 Director General Ken McCallum said artificial intelligence could change the stakes on existing or new bioweapons. He spoke at MI5 headquarters on the River Thames and announced the launch of the AI Threat Assessment Centre. He said the agency has uncovered suspects using AI to improve attacks, including with AI-enabled drones. McCallum also said the UK is devoting more resources to foreign state threats than at any time since the Cold War. He called Russia the main aggressor and noted a fourfold rise in state threat cases over the past decade.

Dalio sees AI bubble, strategist targets zombie companies

Ray Dalio, founder of Bridgewater Associates, said the AI market is close to bursting like a classic bubble. He pointed to rising debt costs and higher interest rates as key risks. In response, strategist Jordi Visser said the real danger lies with zombie companies kept alive by low rates. Visser said AI demand is still strong, citing Lumentum Holdings, whose optical components are sold out through early 2029. He also noted Nvidia trades at about 20 times expected earnings, far below dot-com era levels. Visser warned that rate-dependent sectors like utilities and autos face the biggest bubble risk.

Professor Rob Reich joins Newsom's AI advisory board

Stanford professor Rob Reich was named to Governor Gavin Newsom's AI advisory board. The board was created through an executive order issued in September 2026 to improve AI safety governance. Reich said the group will work on independent assessment and oversight of AI systems. Other members include the first White House Chief Digital Officer and experts in AI alignment and technology policy. Reich, who teaches political science at Stanford, said states like California can serve as laboratories for AI policy. He also referenced the OpenAI Hugging Face attack in July as a real example of AI misalignment risks.

AI speeds up work but better jobs need human choices

A new analysis shows AI helps 63 percent of U.S. workers complete tasks faster, but that does not automatically lead to better jobs. The author, Christos Makridis, said employers must decide how to use the time savings to improve work quality, learning, and career growth. He said companies should provide paid time for all workers to learn and test AI tools, not just volunteers. Workers should also be able to flag errors without being seen as resistant. Makridis argued that turning time savings into better jobs must remain a human-driven effort, not something left entirely to software.

Hollywood studios embrace AI to cut costs and boost creativity

Hollywood is increasingly embracing AI in filmmaking. Producer Ryan Kavanaugh says AI helped cut the cost of the film Bitcoin from 200 million to 70 million dollars. Directors like Steven Soderbergh and Doug Liman have used AI tools in their work. Industry leaders say AI enhances creativity rather than replacing it, even as many Americans worry about job loss.

UK AI Security Institute regains access to frontier models

The White House reversed a decision that had blocked the UK's AI Security Institute from accessing new AI models. Models from Anthropic, Google DeepMind and OpenAI can again reach the agency before launch. The dispute began in September when the Trump administration told companies to stop sharing models with the institute. AI Minister Kanishka Narayan plans to announce tougher reporting rules for AI security incidents.

AI robots now sort returned clothing in European warehouses

Zalando, CEVA Logistics and Sereact have deployed AI robots to handle product returns in Germany and Poland. The robots can sort clothing items they have never seen before using Sereact's Cortex AI platform. UPS-owned Happy Returns improved sorting efficiency by 50 percent using Geekplus robots. A PYMNTS report found 58 percent of merchants expect AI agents to automate returns and refunds.

California law keeps humans in charge of AI medical decisions

California signed Assembly Bill 1979 requiring licensed health professionals to keep control over patient care decisions. The law takes effect January 1 and prevents AI from making medical decisions on its own. Hospitals must ensure doctors can exercise independent judgment when AI tools guide patient care. Clinicians can still use AI tools to inform their decisions.

Goodfire focuses on making AI decision-making clearer

Goodfire CEO Tom McGrath says researchers must learn how AI systems reach their decisions. The company explores mechanistic reasoning to break down complex AI tasks into smaller parts. This approach aims to make AI decisions transparent and explainable. McGrath says the main challenge is applying this method to complex tasks and all types of AI systems.

Why AI needs goals set by society

The article argues that AI development currently lacks clear societal goals. It compares this to car emissions rules, which society sets before engineers build cars. Without agreed-upon limits, AI optimization may ignore democratic values like fairness and human rights. The author urges embedding public interest and human empowerment into AI from the earliest stages of design.

Sources

• AI Poses Grave National Security Risks, Says MI5 Head
• AI could 'change the stakes' on bioweapons: UK spy chief
• Ray Dalio Says the AI Bubble Is Close to Bursting. A Strategist Says Zombie Companies Are the Real Risk
• Professor Rob Reich to serve on Newsom’s AI advisory group
• AI is making work faster, but creation of better jobs won’t be automatic
• Open to AI: Why Hollywood Embraces Artificial Intelligence
• Washington Backs Down: UK’s AI Security Institute Regains Early Access to Frontier Models
• AI Robots Are Ready for Whatever Shoppers Send Back
• California Bans AI From Making Medical Decisions on Its Own
• Researchers Need to Learn How AI Makes Decisions, Says Goodfire’s Tom McGrath
• What Are the Goals for AI?