收听本期播客
阅读正文
Artificial Intelligence (AI) has taken the world by storm, with new advancements promising remarkable capabilities. One of the latest developments is OpenAI’s GPT-5, the upgraded version of the widely used chatbot, ChatGPT. Released in 2025, this cutting-edge model can perform extraordinary tasks, from designing websites to tackling complex scientific problems with the expertise of a PhD holder. However, behind these impressive features lies a growing concern: the enormous amount of energy required to power such technology.
AI systems like ChatGPT rely heavily on electricity to process data and provide responses to users. In 2023, a basic request, such as asking for a recipe, consumed around 2 watt-hours of energy—comparable to running a light bulb for a few minutes. With GPT-5, however, experts estimate that a similar request now demands up to 20 times more energy, sometimes reaching 40 watt-hours for a medium-length answer. This is equivalent to keeping a light bulb on for nearly 40 minutes. The reason for this dramatic increase is thought to be the model’s larger size and greater complexity. Additionally, GPT-5 can process not only text but also images and videos, further driving up its energy needs.
The scale of this energy consumption is staggering. With billions of daily requests on ChatGPT, the total electricity used by GPT-5 could rival the daily power needs of over a million households in the United States. Despite these figures, OpenAI has not released official data on GPT-5’s energy usage, leaving researchers to rely on estimates. Many experts are now urging the company to be more transparent, arguing that the environmental impact of AI must be addressed openly as its role in society expands.
This issue raises important questions about the future of technology. While AI offers endless possibilities, it also places significant demands on resources. Some suggest that efforts should focus on making AI systems more energy-efficient, while others question whether the unchecked growth of such technology should be limited to protect the planet. The debate continues: is the progress of AI worth the environmental cost, or should stricter regulations be put in place to ensure sustainability? As AI becomes more integrated into our lives, finding a balance between innovation and responsibility will be crucial.
