As generative AI continues to change our lives, it’s hard to ignore the growing concerns surrounding its environmental impact. Critics warn of alarming possibilities, including a 99.9% chance of it causing humanity’s downfall and the potential for mass job loss. AI is being embraced in various sectors, from healthcare to education and entertainment. However, this rapid integration raises serious questions.
Elon Musk recently pointed out that we might soon face the world’s biggest technological breakthrough in AI, yet we could run out of electricity to support it by 2025. Even more alarming is the high water demand needed to cool these AI models. Reports from last year revealed that Microsoft Copilot (formerly Bing Chat) and ChatGPT use about one bottle of water for every query. Moreover, a recent study found that OpenAI’s GPT-4 model consumes up to three bottles of water just to generate 100 words.
Also read: Here is why Sam Altman called OpenAI’s o1 model deeply flawed
New findings from an ongoing study titled Making AI Less Thirsty suggest that ChatGPT actually uses up to four times more water for cooling than previously thought, reports The Times (via Windows Central). This study is based on earlier versions of ChatGPT running the GPT-3 model. So, this indicates that newer models like GPT-4 and GPT-4o, with their advanced features, might require even more cooling resources.
This situation is concerning from an environmental standpoint. As AI grows more powerful, meeting its electricity and water demands becomes more challenging. Microsoft aims to be carbon-negative by 2030, but the rising resource consumption from AI makes this goal tougher to achieve.
Also read: OpenAI’s new model might be capable of deceiving and cheating, suggests godfather of AI
Additionally, OpenAI reportedly spends around $700,000 each day to keep ChatGPT operational. With a recent market cap of over $157 billion, the company faces immense pressure to generate profits amid ongoing concerns about sustainability. The need to transition to a for-profit entity within the next two years poses further challenges, as OpenAI must navigate obstacles from investors, regulatory bodies, and its internal board.