Project launch and timeline

• The group aligned on objectives and milestones for the generative AI environmental impact white paper, with Phase One focused on mapping the landscape and producing an interim report by mid‑March 2026.

Building the right expertise

• Plans were put in place to assemble an expert working group spanning carbon accounting, AI deployment, video production and green algorithms, with participants contributing suggestions to fill key knowledge gaps.

Prioritising video generation use cases

• The group explored text‑to‑video and image‑to‑video approaches, agreeing that final prioritisation should reflect real commercial practice and focus on applications where AI genuinely substitutes existing production activities.

Defining meaningful counterfactuals

• Participants emphasised the need for realistic comparisons, positioning AI against modern production alternatives such as virtual production and visual effects, rather than simplistic “AI vs traditional” framings.

Data challenges acknowledged

• Significant gaps remain in high‑quality, comparable data for video and audio generation. The group agreed the work should focus on orders of magnitude and relative differences, supported by a mix of primary partnerships and existing research.

Scope beyond video noted

• While video generation will be the primary focus for quantitative analysis in Phase One, other GenAI applications - including text, image and audio use cases - will be captured in the literature review to inform future research.

Strong commitment to collaboration

• Participants volunteered to support expert review, data sourcing and amplification of draft outputs, reinforcing the shared ambition to develop transparent, decision‑useful guidance.

The session set a clear direction for the work ahead, combining methodological rigour with real‑world relevance.