SUMMARY
This fifth episode of the podcast “Responsible Digital Technology Explained,” presented by The Green IT Day, is titled“Can Artificial Intelligence Be Responsible?” It features Magali Germond, founder of LumIAire Conseils, a company that supports executives in integrating trustworthy artificial intelligence (AI) through consulting, training, and webinars.
She defines responsible AI as a process rather than a static state: ethical AI does not exist, but it is possible to do better by minimizing negative impacts and maximizing positive ones. This requires the skills to make informed choices. Since technology is not neutral, the analysis must cover the full spectrum of impacts.The environment is affected beyond mere usage, with an environmental footprint linked to the design of generative AI, its training, and its operation (a generative query is estimated to consume about ten times more energy than a Google search). However, this is not the only impact of using AI, as we must also consider the effects on work organization, corporate governance, fundamental rights, and the brain. Agent-based AI, which coordinates multiple agents, further increases energy consumption.
In practice, Magali Germond cautions against the tendency to rely on AI out of habit. We must systematically ask ourselves about the added value of its use and avoid resorting to AI when the task can be performed without it. Finally, she recommends querying generative AI only on topics for which we already know the answer, so that we can challenge it.
Keywords: acculturation, discernment, carbon footprint, critical thinking, Green IT, artificial intelligence, digital restraint, ecological transition.
Illustration: The Green IT Day
LISTEN - PODCAST
