Blog

Professor Geoffrey Hinton’s warning about AI’s kill switch highlights a difficult governance challenge, but stronger controls, testing and cooperation could still help shape a safer future.

Artificial intelligence is beginning to reshape air traffic management, but a new US safety debate highlights the challenge of balancing operational efficiency with accountability.

Britain’s intelligence services are creating a clearer picture of AI-driven threats, as autonomous systems, cyberattacks and potential biological risks reshape national security.

NVIDIA’s OpenShell arrives after a landmark AI-agent breach, promising a new security architecture designed to contain autonomous systems before they cross critical boundaries without authorisation.

An OpenAI agent accessed non-public Australian government files, raising a bigger question: what happens when autonomous AI systems encounter digital barriers designed to stop them?

A bitter dispute between mathematicians and OpenAI raises urgent questions about data, credit, competition and who controls discovery in the age of AI agents.