Jessica Smith

As generative AI reshapes how we search, consume and interpret information online, the need for reliable attribution and effective labelling is more urgent than ever. At Ofcom, Jess is leading research into how users respond to AI-generated content - from watermarks and metadata to annotations like Community Notes - and how these measures influence trust, interpretation, and onward sharing. Drawing on insights from Ofcom’s Deepfake Defences 2 report, this session explores what’s working, what’s not, and what platforms, policymakers and users need to know.

• Evaluating watermarking and metadata: how robust are they to basic manipulation?
• Analysing user response to AI labels and annotations across different content formats
• Understanding how chatbots and GenAI are shaping online search and fact-checking behaviour