AI needs "rigorous" testing and safeguards to contain risk, Andrew Bailey says.

Read the complete, original reporting at BBC Technology for the full story, additional context and any updates.