UK intelligence chiefs warn AI poses major national security threat

British intelligence leaders have warned the government that artificial intelligence poses a significant national security threat. Heads of MI5, MI6, and GCHQ reported to the cabinet that hostile states, including Russia and Iran, as well as terrorist groups, could utilize AI for cyberattacks, election manipulation, and weapon research.

Recent government hackathons revealed substantial vulnerabilities, with 407 flaws discovered across nine agencies. One critical vulnerability allowed for the potential alteration of computer code to impact essential digital services, highlighting how low-cost AI-driven attacks could cause major damage.

In response to these risks, the United Kingdom is positioning itself as a leader in establishing international AI safety standards. The UK AI Safety Institute is actively testing advanced models from global companies like Google to identify risks related to biochemical information and safety bypasses. Microsoft has also identified the UK as a suitable nation to lead these safety criteria due to its specialized testing institutions and strong ties with the US government.

South Korea is seeking to participate in this international safety framework, potentially acting as a third-party evaluator between the US and China. However, experts note that South Korea faces significant gaps compared to the UK in terms of specialized personnel, computing resources, and budget.