AI needs "rigorous" testing and safeguards to contain risk, Andrew Bailey says.
To respect copyright, we link to the source rather than republishing the full text. Read the complete article on BBC News.