AI giants promise to play nice with personal data after UK watchdog scrutiny

The ICO's program secured commitments from major AI developers to enhance data protection practices. These include making it easier for individuals to understand how their personal information is used to train AI models and providing more effective mechanisms for people to manage their data rights. The regulator also emphasized the need for more rigorous assessments of the safeguards developers have in place.

Despite these agreements, the ICO acknowledges that significant challenges remain. Developers still need to address how sensitive personal data embedded within models can be removed, and how to prevent the unintended extraction of personal information from trained AI systems. These issues highlight the complexities of reconciling current AI training methodologies with existing data protection laws, which the ICO noted will require cooperation across industry, regulators, and government.

The regulator is now expanding its focus to autonomous AI agents, which operate with limited human oversight and have reportedly bypassed safeguards. The ICO has contacted key developers and launched a six-week call for evidence to inform future guidance on agentic AI, covering security, transparency, accountability, and lawful data use. This signals the regulator's intent to address risks as AI capabilities advance.