Will sustainable AI ever be a reality?

In this episode, Oliver Cronk is joined by colleagues James Camilleri and Hélène Sauvé to explore the complex topic of AI sustainability. It's a conversation that spans the environmental impact of AI, from carbon emissions to water usage, and examines whether the industry narrative matches reality. They delve into the trade-offs between energy and water, the role of lifecycle analysis, and how approaches such as federated learning and smaller models could help. They also ask whether "green AI" is more than marketing hype, and what practical steps organisations can take to make AI genuinely sustainable.

Useful links for this episode

• Technology Carbon Standard – Open-sourced by Scott Logic
• Balancing AI Innovation and Sustainability: Our presentation at HM Treasury ID25 – Oliver Cronk
• Greener AI - what matters, what helps, and what we still do not know – James Camilleri
• Introducing AI-related emissions into the Technology Carbon Standard – Hélène Sauvé
• Mapping the carbon footprint of digital content – Hélène Sauvé
• Sustainability in AI - why we must listen to the smaller players – Hélène Sauvé
• Sustainability topic on the Scott Logic Blog
• Beyond the Hype: Can technology sustainability really make a difference?
• Beyond the Hype: Can we do better than 'carbon aware' computing?
• Artificial intelligence threatens to raid the water reserves of Europe's driest regions – Marianne Gros and Leonie Cater, Politico
• Mistral
• Green Tech South West
• Will new models like DeepSeek reduce the direct environmental footprint of AI? – Chris Adams
• Government Digital Sustainability Alliance (GDSA)
• Hugging Face
• Empire of AI – Wikipedia
• Computer Says Maybe podcast
• GreenIO podcast
• V-Bucks – Fortnite Wiki