Findings from an expert workshop exploring the issues arising from AI companies’ use of news as data for training and grounding AI systems. The rapid development and productisation of generative AI, often using news as data for training and grounding models, has prompted the news industry to respond – some with lawsuits, others with data licensing agreements, but all with renewed attention to their position in a shifting media ecosystem. As the UK government considers expanding the text and data mining exception in copyright law and promotes AI adoption and industry in the AI Opportunities Action Plan, there is a need to understand how changes wrought by generative AI are impacting the news media specifically. In November 2024, BRAID (Bridging Responsible AI Divides) and the Ada Lovelace Institute organised an expert workshop, under Chatham House rules, to explore the state of play in news publishing, including pressing challenges, current responses, and emerging safeguards and solutions. We heard from major news publishers, AI and natural language processing experts, lawyers and legal scholars, and copyright experts. The workshop sought to answer three core questions: 1. What is the state of play in news publishing regarding news as data for AI models?2. What concerns does this situation raise for news publishers and society?3. What solutions/safeguards could support news publishers and the sustainable provision of a plurality of public interest news?