Photo by Alena Koval on Pexels

While AI companies like OpenAI and Google are providing some data on energy consumption, significant uncertainties remain regarding the true environmental cost of artificial intelligence. Reported figures, such as OpenAI’s claim that a ChatGPT query uses 0.34 watt-hours and Google’s estimate of 0.24 watt-hours for a Gemini query, lack crucial details. Experts argue for more comprehensive data encompassing various AI modalities, including energy usage for resource-intensive tasks like video and image generation. Current data primarily reflects chatbot interactions, representing a small portion of AI’s expanding applications. The proliferation of data centers powering AI is a major concern, with projections suggesting AI could consume the equivalent of 22% of US household electricity by 2028. Microsoft’s rising emissions, attributed to AI development, underscores the challenge of balancing innovation with environmental responsibility. The question remains whether AI’s potential efficiencies can offset its substantial energy demands, casting doubt on the long-term sustainability of the AI revolution.