
RG Lab
RG Lab is France Televisions' broadcast-technology research-and-development unit that uses Roland-Garros tennis as a live production sandbox for AI and immersive media experiments.
Last refreshed: 17 June 2026 · Appears in 1 active topic
France TV named its AI vendors at Roland-Garros; why won't more broadcasters do the same?
Timeline for RG Lab
Disclosed the full AI stack France Televisions ran at Roland-Garros 2026, naming three commercial vendors
Media's AI Pivot: France TV names its AI vendorsBackground
RG Lab is France Televisions' broadcast-technology research-and-development unit, using the Roland-Garros tennis tournament as a live production sandbox for AI and immersive media experiments. On 11 June 2026 France Televisions disclosed the AI stack RG Lab deployed at Roland-Garros 2026, making it the first European public broadcaster to name its generative-AI vendors in an active live-sports production context. The disclosed stack included Decart's Lucy 2.0 for real-time video transformation, Gracia AI for 4D Gaussian Splatting volumetric video, Topaz Labs' Starlight Precise 2.5 for video remastering, an AI virtual mirror for interviewee anonymisation, agentic AI for an Android XR application, and live audience-chat analysis.
France Televisions is France's public broadcasting group, encompassing France 2, France 3, France 4, France 5, and France 24. RG Lab is its experimental unit specifically linked to the Roland-Garros clay-court season, giving it a recurring annual production window in which to test and iterate AI tooling under real broadcast conditions. The lab's vendor-disclosure model is the inverse of Netflix's approach on The Eternaut, where Runway's involvement was not credited until a leak, and it positions France Televisions as a transparency exemplar among broadcasters adopting generative AI.
The RG Lab approach is significant because it produces auditable evidence of what broadcasters can actually deploy in live sports production, distinct from promotional announcements. The 2026 stack spans video latency, volumetric capture, remastering, privacy, app development, and audience interaction, covering most major categories of generative-AI use in live broadcasting.