The Climate Question Q+A: The carbon footprint of your tech and AI
19 snips
May 10, 2026 
Caroline Steel, BBC CrowdScience presenter who studies estuary ecology and mitigation trials. Akshat Rathi, Bloomberg climate reporter exploring digital carbon and energy trends. Justin Rowlatt, BBC Climate Editor with analysis on climate impacts and policy. They discuss streaming and device manufacturing emissions. They dig into AI, data centre energy and local impacts. They explore estuary vulnerability and a 'fish disco' mitigation trial.
AI Snips
Chapters
Transcript
Episode notes
Making Phones Causes Most Emissions
- Production often dominates the carbon footprint of battery-powered devices like phones, with manufacture ~50 kg CO2e per smartphone.
- Lifetime use emits only about 8 kg CO2e, so extending device life yields big gains.
TVs Emit More When Used Than Made
- For mains-powered devices like TVs the balance flips: production (~200 kg CO2e) is substantial but usage (~600 kg CO2e) dominates.
- So device type determines whether making or using drives emissions.
AI Searches Use Little Energy Individually
- Individual AI searches have very low carbon footprints; data centres and aggregate AI energy use matter more.
- Estimates place AI at ~0.5% of world electricity (≈0.1–0.2% total energy), so scale is limited currently.