Consultation on AI & web-scraped data

Generative AI means AI that can create content such as text, audio and imagery. Typically, these AI models are trained on extensive datasets.

The ICO wants to hear lawyers’ views on: the appropriate lawful basis for training generative AI models; the purpose limitation principle (that personal data should only be used for specified purposes) in the context of generative AI; expectations around compliance with the accuracy principle (requirement to take reasonable steps to erase inaccurate data); and expectations on compliance with data subject rights.

In coming months, the ICO will set out its emerging thinking on these questions, and is inviting legal advisors and interested parties to share their views. The first chapter will cover the lawful basis for training AI models on web-scraped data, and is open until 1 March.