A new study reveals OpenAI's dated GPT-3 uses more water than previously assumed — ChatGPT needs up to four times more water to quench its thirst for cooling data centers
By Kevin Okemwa,
1 days ago
What you need to know
A new study reveals that OpenAI's ChatGPT running GPT-3 uses up to four times more water than previously thought.
A separate study indicated that OpenAI's GPT-4 model (referred to as mildly embarrassing at best) uses up to 3 water bottles to generate a mere 100 words.
Generative AI's high demand for electricity and water is subtly hurting large tech corporations' sustainability and environmental goals.
Despite the rapid adoption and advancements in the generative AI landscape , users have raised critical concerns about the technology, including a 99.9% probability it'll end humanity , take over jobs from humans , and more. Major tech corporations like OpenAI and Microsoft are at the frontline in the AI landscape with their flagship AI models have revolutionized our lives.
However, it's still unclear whether the advantages of broad access to AI outweigh the cons. While the technology is being integrated into workflows and even across medicine , education , computing, entertainment, and more, it raises critical concerns equally.
In 2020, the University of Riverside published a study indicating that ChatGPT consumes approximately 2 liters of water to generate 50 responses to queries. Now, we learn that the technology potentially consumes more water than we thought.
According to an ongoing study dubbed Making AI Less Thirsty, slated to be published in Communications of the ACM, OpenAI's ChatGPT consumes up to four times more water for cooling than previously thought (via The Times ). For context, the study is based on an earlier version of ChatGPT running OpenAI's GPT-3 model.
As you may know, OpenAI has shipped new next-gen AI models since GPT-3, including GPT-4 and GPT-4o which spot advanced capabilities like real-time audio, text, and vision capabilities. GPT-3 is far less capable when compared to OpenAI's flagship models . Based on this premise, the firm's newer AI models may use more water for cooling because of their advanced capabilities to handle more complicated tasks.
Not a good look on the environmental and sustainability front
Major tech corporations invested in the AI landscape, such as OpenAI, Microsoft, Google, and others, are trapped between a rock and a hard place. As AI becomes more advanced and powerful, it's becoming increasingly difficult to meet its high electricity and water demands.
Aside from the dire need for resources to support their ventures, exploration of the technology acts as a major deterrent to achieving their sustainability and environmental goals. In case you missed it, Microsoft is working towards becoming carbon-negative by 2030 .
According to a recent sustainability report, Microsoft reported 22.5% of total water consumption, while Google and Meta used 17% each. It remains unclear how these companies will manage AI's high demand for cooling water as models become more sophisticated.
However, the round funding features a critical threshold that could leave OpenAI in a worse position . OpenAI reportedly needs to turn into a for-profit entity within the next two years or run the risk of returning the money raised by investors. Experts speculate that the transition might encounter several roadblocks from OpenAI's board of directors, staff members, the government, and regulatory institutions.
We at Windows Central are scouring the internet for the best Prime Day deals and anti-Prime Day deals , but there are plenty more discounts going on now. Here's where to find more savings:
Get updates delivered to you daily. Free and customizable.
It’s essential to note our commitment to transparency:
Our Terms of Use acknowledge that our services may not always be error-free, and our Community Standards emphasize our discretion in enforcing policies. As a platform hosting over 100,000 pieces of content published daily, we cannot pre-vet content, but we strive to foster a dynamic environment for free expression and robust discourse through safety guardrails of human and AI moderation.
Comments / 0