# MI5 Launches AI Threat Centre Amid Escalating Security Concerns UK intelligence establishes AITAC as industry leaders warn of potential future AI incidents. - **Source:** [The Next Web](https://brifing.si/source/afe24b20-b275-4c19-b9a6-8bb3c8babae7) - **Published:** 2026-10-16 - **Topic:** [Business](https://brifing.si/c/business) - **Original article:** - **On Brifing:** https://brifing.si/s/46095c18-2a1c-4332-abec-7531e725a3b0 ## MI5 Establishes AITAC Unit MI5 director general Sir Ken McCallum announced the creation of the AI Threat Assessment Centre, or AITAC, during his annual update on October 13. Located at Thames House in London, this unit integrates experts from the AI Security Institute, the National Crime Agency, and the National Cyber Security Centre to monitor emerging risks. ## Centralizing Intelligence Operations AITAC functions as MI5’s third specialized center, following existing units focused on terrorism and state threats. McCallum stated the center aims to provide the government with a single version of the truth. This centralized approach prevents conflicting departmental views and helps separate realistic threats from speculative noise in the rapidly evolving AI landscape. ## Rising State-Sponsored Threats MI5 reports that its state-threat caseload has quadrupled over the last decade, requiring more resources than during the Cold War. The agency now classifies national resilience, termed Protect, as a core mission alongside counter-terrorism. Officials warn that foreign actors are actively spying on tech startups, necessitating urgent improvements in corporate cybersecurity protocols. ## AI-Enabled Reconnaissance Risks Intelligence officials have already identified terrorists attempting to exploit AI to refine their attack methodologies and reconnaissance capabilities. Furthermore, the integration of AI into drone technology introduces new tactical challenges. McCallum expressed particular concern regarding the potential for AI to lower barriers for developing or enhancing dangerous bioweapons, despite current technical limitations. ## OpenAI Testimony to Parliament During a House of Commons hearing, OpenAI’s Tom Duff Gordon admitted that the company has encountered multiple security incidents involving AI agents. He could not guarantee that reported cases were exhaustive, noting that OpenAI previously paused reinforcement-learning training on advanced models following a July incident involving Hugging Face to ensure safety. ## Internal Industry Criticisms Former researchers from major AI labs have raised concerns regarding internal safety cultures. Daniel Kokotajlo, who left OpenAI, cited a lack of responsible behavior, while Steven Adler noted internal resistance to labeling technologies as dangerous. These testimonies highlight ongoing tensions between rapid commercial development and the implementation of robust, transparent safety guardrails. ## Expert Perspectives on Misuse Dr. Filippa Lentzos of King’s College London emphasizes that while AI amplifies existing expertise, significant barriers like materials and delivery mechanisms remain for bad actors. Prof. Tom Inglesby of Johns Hopkins noted that no country currently possesses formal policies to address the specific risks of AI-enabled bioweapon development or misuse. ## Legislative Inquiry and Oversight The Business, Innovation, Science and Trade Committee has launched a formal inquiry into AI safety. Chair Liam Byrne plans to publish a special report ahead of potential amendments to the Cyber Security Bill. This legislative push reflects growing parliamentary pressure to ensure that voluntary industry safeguards keep pace with accelerating AI capabilities. ## Sources covering this story - https://thenextweb.com/news/mi5-ai-threat-assessment-centre-openai-mps — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._