- 16:15Mozambique Police Clash with Protesters Over Election Disputes
- 15:35Trudeau Government Orders TikTok to Close Canadian Offices, but Access Remains Unrestricted for Users
- 14:55Israeli Parliament Approves Controversial Law to Deport Relatives of Alleged ‘Terrorists
- 14:20MP Mike Amesbury Faces Assault Charge Following Alleged Street Incident
- 13:12Trump's Election Sparks Unprecedented Surge in Billionaire Wealth
- 12:12Morocco Showcase Summit: A Gateway to Tourism and Investment Opportunities
- 11:13Urgent Calls and Delayed Action: A Closer Look at Valencia's Flood Response Crisis
- 10:40Over a Decade of Service to Casablanca's Mobility
- 09:57Trump's Return to Power: Promises and Challenges Ahead
Follow us on Facebook
The Alarming Ecological Impact of Conversational AI: A Ticking Time Bomb?
As conversational artificial intelligence systems like ChatGPT gain popularity, their environmental footprint raises serious concerns. Resource-intensive and water-hungry, these groundbreaking technologies could become a genuine ecological time bomb if measures are not swiftly taken.
Chilling Figures
Alex de Vries, an employee of the Central Bank of the Netherlands, sounded the alarm on the Digiconomist website. His calculations reveal that if Google's chatbot were used for every web search, its annual electricity consumption would reach an astronomical level of 29 billion kilowatt-hours, equivalent to the consumption of countries like Kenya, Guatemala, and Croatia combined.
The meteoric success of ChatGPT, now handling 200 million daily queries, underscores the urgency of the situation. Estimates suggest that this AI system consumes over half a million kilowatt-hours of electricity, 17,000 times more than an average American household. Such a stark realization prompted Sam Altman, CEO of OpenAI, to openly question the sustainability of such technologies at the recent Davos Forum.
Necessary Global Awareness
While the digital industry may not be the most polluting, the impact of AI systems could quickly be felt, as the internet already accounts for nearly 4% of global greenhouse gas emissions. Furthermore, a study by American researchers revealed that training GPT-3 in Microsoft's data centers required 700,000 liters of freshwater, equivalent to the amount needed to cool a nuclear reactor.
Faced with these major environmental challenges, global awareness is imperative. Experts like Martin Bouchard, co-founder of the Canadian data center company QScale, suggest that computing capacities for search engines like Bing and Google will need to be multiplied by 4 or 5. A major technological breakthrough in the energy sector, whether in nuclear or renewable energies, seems indispensable to meet the growing demands of these promising yet resource-intensive technologies.