Supercomputers are guzzling water

Good morning, and welcome to a new week of fresh and crispy AI updates! Let’s dive in. 😁

🔨 AI TOOLS OF THE DAY

📝 Dewey: An AI-powered accountability buddy that helps users stay organized and productive. → check it out

🍽️ Unreal Meal: AI-generated images of non-existent meals for recipe ideas and creative projects. → check it out

👀 DuckDuckGoose: A deep learning network designed to detect and recognize manipulated faces in images and videos, including deepfakes. → check it out

AI COMPUTERS ARE GUZZLING WATER 😳

Extra salty French fries make Fry Guy extra thirsty, leading to massive water consumption. AI’s salty supercomputers are thirsty too, but for different reasons! 💧

What’s happening? Microsoft's most recent environmental report disclosed a significant increase in the company’s worldwide water usage, surging by over 30% from 2021 to 2022. This quantity is equivalent to filling approximately 2,500 Olympic-sized swimming pools, totaling nearly 1.7 billion gallons.

Why the water problem? Developing a large language model (LLM) for analyzing human-written text demands a significant energy input, which generates substantial heating in the supercomputers which power this technology. To mitigate this heat, data centers resort to water cooling systems, which can guzzle billions of gallons of water each year. It’s been predicted, for instance, that ChatGPT chugs around 16 ounces of water every time it’s asked a series of between 5 to 20 prompts or questions by a single user.

What is being done? People should not be deceived by thinking major AI companies are unaware of this problem. As if it weren’t obvious, OpenAI said in a statement, “We recognize training large models can be energy and water-intensive.” On top of that, Microsoft said it was investing in research to measure AI’s energy and carbon footprint: “We will continue monitoring our emissions, accelerating progress while increasing our use of clean energy to power data centers, purchasing renewable energy, and [make] other efforts to meet our sustainability goals of being carbon negative, water positive and zero waste by 2030.”

AMAZON’S NEW DISCLOSURE POLICY FOR AI-GENERATED BOOKS 📖

Q: Why did the bookworms breakup?

A: Because they were not on the same page. 📚

What’s new? Under pressure from various organizations and authors, Amazon implemented a new policy which now requires authors who wish to sell books containing AI-generated material through its e-book program to declare it in advance.

What does this mean? By distinguishing between AI-assisted content and AI-generated work, Amazon aims to protect consumers from unknowingly purchasing AI-generated books, which could potentially overshadow traditional literary works. AI-assisted content will not have to be disclosed, but AI-generated work will require labeling, as not to deceive customers and damage the reputation of human authors.

What’s next? This policy change by Amazon is a solid first step to fixing the problem of deceptive AI authorship. In a statement on their website, the Author’s Guild thanked “the Amazon team for taking our concerns into account and enacting this important step toward ensuring transparency and accountability for AI-generated content.”

UNITED NATIONS CALLS FOR AI REGULATION IN EDUCATION 🎓

More talking and less doing … that’s how things get done. Right? 😅

What’s going on? The United Nations Educational, Scientific and Cultural Organization (UNESCO) has called upon governments worldwide to implement regulations governing the use of AI in educational and research settings.

What are the recommendations? Key recommendations include the protection of data privacy, potential age limits for classroom use, and ethical requirements for GenAI providers, underscoring the importance of educational institutions validating AI systems for student use.

What does this actually mean? UNESCO recognizes that AI offers tremendous developmental opportunities, but the organization says “it can also perpetuate harm and prejudice.” It’s not clear what the organization means by this or why they believe this is the case. It’s also not clear what they are actually recommending be done about AI in education, aside from the briefing of certain broad regulatory ideas. It seems this is just a way for UNESCO, like many organizations, to claim they are doing something to promote the safe use of AI with no substantial plan on how to do so.

MEME MONDAY 🤣

HAS AI REACHED SINGULARITY? CHECK OUT THE FRY METER BELOW