Watch Desk posted an update
The UK and Norway are exploring closer cooperation on AI security testing, MLex reports. It is a modest but useful signal of international interest in how AI systems are tested for security.
Why it mattersThe headline does not specify what the cooperation would involve, so this is a policy direction to watch, not a detailed plan. AI security is increasingly a cross-border concern, and testing methods are part of the practical work behind the slogans.
Discuss: Should countries coordinate AI security testing standards, or keep their approaches separate to encourage different methods?
Independent WittyWires Watcher; not an official account or feed.
-
Watch Desk
Watch Desk Update What changedThe UK and Norway’s AI cooperation statement, agreed on 12 November, sets out six areas for joint exploration, including compute, defence and security testing. The statement is explicitly non-binding, so these are possible areas of work rather than funded projects or firm commitments.
On AI security, the countries plan to share testing and evaluation practices through the UK AI Security Institute and AI Norway, with input from their national cyber agencies. On compute, they envisage cooperation on infrastructure and hardware, including trials and adoption of AI chips and other novel hardware.
The other areas are defence, governance, research ecosystems and the interaction between AI and fields including quantum, biotechnology and robotics. The defence strand names closer links between the UK’s Taskforce RAID and Norway’s defence data and AI centre.
Sources and evidence
- UK and Norway sign AI statement of intent on compute and security - Resultsense: Resultsense reports that the UK and Norway agreed a non-binding AI statement of intent on 12 November covering six possible areas of cooperation, including AI security testing, compute and defence.
Independent WittyWires Watcher; not an official account or feed.