Copyright Clearance Center (CCC)

Across financial services and insurance companies, licensed research reports and data feeds are being pasted into AI tools that compliance has never reviewed. Each of those prompts can create an unauthorized copy, and standard content licenses rarely extend to AI use. This whitepaper maps where copyright exposure sits across the AI lifecycle and how to contain it.

Use this guide to map where copyright exposure sits across the AI lifecycle and how to contain it. Learn: 🚫Why tuning models on analyst reports and data feeds can breach licenses already in place 📚How RAG systems turn licensed research repositories into infringement exposure ⚖ What three conflicting 2025 fair use rulings mean for firms relying on licensed databases 👀 How employee surveys pinpoint the teams carrying the most copyright risk 🚦Which policy and training updates close the gap between sanctioned and shadow AI use

Find where shadow AI is creating copyright exposure inside your firm.

How do today's publishing leaders guide AI adoption and strengthen oversight? Join governance and leadership advisor Mari Wallgren of LeadMWell, in Copyright Clearance Center (CCC)'s upcoming webcast and gain valuable insights on aligning AI governance in publishing with organizational values. Register at https://lnkd.in/eE2FTA6F

If public data is freely available online, can it be used in an AI system? Remember: much of what appears online is protected by copyright, even if it is freely accessible. Free access does not automatically grant you reuse rights for copying, training, or other AI use. Web scraping, downloading, ingesting online content into an AI system creates copies that may infringe unless you have permission or a valid legal basis for doing so. If a piece of work is in the public domain, or if a piece is available under open licenses that actually allow this kind of reuse, these could be instances where this type of use is acceptable. But you need to verify that first before using it. The big takeaway? Always check the terms of use or rely on trusted licensed sources before using online material in any AI workflow.

"A lot of the licenses banks rely on, for market data and other third-party proprietary content, never contemplated content being ingested into an AI system. That is true even for a closed instance that is not retraining a public model like Copilot but is retraining your own system. None of that was contemplated in the original agreements, so there is a real question of how you redo or manage all of it. And in a company with hundreds of thousands of people, even with governance, policies, and education in place, you cannot always know the full scale at which these tools are being used." Read our full Q&A with Wells Fargo's Paul Mussell for a look inside the company's AI risk strategy: https://lnkd.in/gWXVvzEt