Clear project focus agreed

• The group confirmed that the initial scope will concentrate on energy use and related carbon emissions, recognising these as the most material and measurable impacts across media AI workflows.

Media‑specific lens prioritised

• The work will focus on AI used in media and content production, such as video generation and related processes, rather than general enterprise or office AI use.

Key research questions defined

• Participants refined the core questions the project will address, including how best to define functional units, set system boundaries, identify credible counterfactuals and estimate emissions ranges across different use cases.

Structured two‑phase approach

• The project will progress from a literature review and methodological groundwork into primary data collection and detailed use‑case modelling.

Independent and research‑led delivery

• Roles were clarified to ensure independent editorial control alongside academic input, strengthening the credibility and robustness of the outputs.

Flexible participation model

• Members can engage at different levels, from discussion and draft review to contributing anonymised or attributable data and case studies - without any requirement to sponsor the work.

Momentum and next steps

• Early actions include recruiting expert input, sharing relevant research, and identifying primary data partners to support Phase Two analysis.

The session established a strong foundation for collaborative, evidence‑based work on the environmental impacts of generative AI, with a shared ambition to deliver practical, decision‑useful insights for the media sector.