The UK has an AI Security Institute involved in testing AI models' cybersecurity behavior.
Sources
- Anthropic's AI used fake identities, malware in rogue ... - Ars Technica (arstechnica.com)
- https://www.aisi.gov.uk/blog (aisi.gov.uk)
- https://www.engadget.com/2230628/openai-anthropic-models-hacking-spree-test-uk-ai-research-institute/ (engadget.com)
- https://dataconomy.com/2026/08/04/uk-ai-security-institute-unsanctioned-actions-online/ (dataconomy.com)
More in Technology
- Each Lunar Orbiter carried two cameras—one high-resolution and one medium-resolution—and used 70-millimeter film rather than standard 35-millimeter film.
- Apple's iTunes U application launched in January 2012.
- Ars Technica published an article titled 'Spectre, Meltdown researchers unveil 7 more speculative execution attacks' on November 14, 2018.
- Google researchers labeled gender categories in their smile-detection model as 'Gender 1' and 'Gender 2' specifically to reduce the risk of introducing unconscious societal bias.
- The publication date of 'Harmonising Artificial Intelligence' is December 2021.
- Patricia Shaw is an IEEE CertifAIEd AI ethics assessor.