Data centre (illustrative) for the ECHO AI webinar on the environmental cost of AI
Large AI models use energy, water and hardware, from training through to everyday use. That cost, often hard to see, is the subject of the webinar “The hidden environmental cost of AI development”, run by ECHO AI on 16 December 2026, at 18:00, on Google Meet.
The session looks at the carbon footprint of these systems and at the Jevons paradox, which the organiser puts on the table: when each operation becomes more efficient, total consumption can still rise. The event page suggests reading papers on arXiv about the energy, water and emissions involved in training and using language models. From there it turns to a practical question. What would a realistic rule for responsible AI use look like in a Romanian university or organisation?
The meeting is online, on Romania time. To receive the Google Meet link, you sign in on the event page and register your interest. ECHO AI has not published an end time or a fee.