A new report from The Alan Turing Institute has called for focus on the national security risks posed by adversaries co-operating on AI.

For the report, which was published today by the Institute’s Centre for Emerging Technology and Security (CETaS), researchers explored the current global AI landscape including cooperation between key players.

The research examined each “CRINK” state (China, Russia, Iran and North Korea) and AI-related interactions to identify future avenues for co-operation on hostile AI.

It cites China as one of the most capable states globally, maximising its relationship with Russia to signal intent and innovation. China also appears to be actively expanding security and technological cooperation beyond CRINK states.

The report also considers the potential implications on UK national security by adversaries – including the sharing of hostile AI capabilities in the fields of information warfare, offensive cyber operations, and military tech.

The study involved a review of academic and “grey” literature, open-source investigation to identify case studies of collaborations, and a scenario generation exercise within government and academic experts.

Recommended reading

• Report: Appetite for AI Investment Higher Than Manageable for CIOs
• Most UK Businesses Plan to Increase AI Spend Despite Lack of ROI
• Report: Only 4% of Firms Turn AI Investment into Scalable Value

The new report found that although there’s no direct evidence of multilateral CRINK co-operation on AI, there are “early indications” of bilateral co-operation in this area.

In order to protect UK national security, it’s been suggested that the government prioritise disrupting hostile AI cooperation, improve the UK’s resilience, and build trusted partnerships.

Future work should also be done to identify vulnerabilities, establish unacceptable red lines in hostile AI co-operation, and monitor early signs of hostile AI co-operation such as talent exchange programmes, it’s been said.

“The extent to which adversaries and geostrategic competitors can acquire, develop, apply and export AI is of significant strategic interest to the UK and its allies,” explained Megan Hughes, lead author and research associate at CETaS.

“This research provides evidence on contemporary bilateral and multilateral cooperation on hostile AI and more importantly, UK policy priorities to mitigate against future risks to our national security.”