Gracenote discussed the growing challenges around artificial intelligence and copyright at the FICCI Frames panel ‘The Great Content Race: Artificial Intelligence and Copyright’. The discussion focused on the need to identify, understand and attribute content used by AI systems.

According to a Gracenote press release today, the panel featured Uday Singh of the Motion Picture Association, Kaushik Moitra of Bharucha and Partners, Manasvi Sharma of Gracenote and Anil Lale of JioStar. The discussion was moderated by Mihir Rale.

As AI becomes increasingly embedded across the media and entertainment industry, the discussion also covered questions around how existing copyright frameworks should apply to generative AI.

A recent Department for Promotion of Industry and Internal Trade (DPIIT) discussion paper examined issues around the use of copyrighted material for AI training, copyright and authorship of AI-generated works, moral rights and potential liability for infringement.

Gracenote said content identification is increasingly moving beyond complete programmes and titles to scenes, moments and portions of a stream. If identity can persist at that level, clips, scenes or other content derived from an original programme can potentially be associated back with the underlying programme.

Gracenote also highlighted the importance of content provenance and consistent identifiers for tracking content across its lifecycle.

These mechanisms could provide greater visibility into what content has been used and how it has contributed to AI systems and outputs, supporting discussions around licensing and remuneration.

Gracenote also cited its recent research, which found that ungrounded AI models fabricated every tested metadata attribute for nearly one in five movie and TV titles examined.

The company said this highlights the importance of authoritative content data in AI-powered discovery experiences.