AI in the News: Hype vs. Reality, and the Future of Chatbot Accuracy

Hi! Here’s your Tuesday, September 24, 2024, edition of AI in the News. I hesitated to call it a ‘Special Skepticism Edition’—there are a lot of nuances and interesting thoughts to explore!

Generative AI Hype Feels Inescapable. Tackle It Head-On With Education - Wired

A book I’ve been looking forward to for a while! Arvind Narayanan and Sayash Kapoor critique the hype around AI in their book AI Snake Oil.

• They emphasize that their message is not against AI itself but against misleading claims about it.
• The book identifies three main groups perpetuating the hype: the companies selling AI, researchers studying AI, and journalists covering AI.
• The authors turn a skeptical eye toward companies primarily focused on existential risks, like artificial general intelligence.

They also tackle something I’m obsessed with: images accompanying news articles, such as photos of robots or illustrations of altered human brains brimming with computer circuitry. Word! ;)

Will A.I. Be a Bust? A Wall Street Skeptic Rings the Alarm - The New York Times

• Jim Covello, head of stock research at Goldman Sachs, expresses skepticism about the A.I. boom, warning that “overbuilding things the world doesn’t have use for… typically ends badly.”
• He questions whether businesses will see sufficient returns on the projected $1 trillion A.I. spending.
• Covello’s research paper highlights the high costs and shortcomings of generative A.I., stating that it often makes mistakes and is not “smart enough to make employees smarter.”

Is Math the Path to Chatbots That Don’t Make Stuff Up? - The New York Times

• Researchers are developing A.I. systems, like Harmonic’s Aristotle, that can verify their own mathematical answers, reducing the issue of “hallucination” seen in chatbots.
• Harmonic’s founders aim to create an A.I. that can eventually outperform human mathematicians.
• While these A.I. systems can verify mathematical truths, experts caution that they may struggle with the complexities of real-life situations. As noted by Angela Fan: “Once you step out of the mathematical realm, things are very different.”

See also: Sam Altman: AI’s next great leaps forward. “Once it can start to prove unproven mathematical theorems, do we really still want to debate: ‘Oh, but it’s just predicting the next token?’” Altman said.

Energy

Big Tech Is Rushing to Find Clean Power to Fuel AI’s Insatiable Appetite - The Wall Street Journal

Following the article I mentioned yesterday about the rising importance of nuclear energy for tech companies, there’s another interesting piece today:

• Tech giants like Google, Microsoft, and Amazon are urgently seeking clean energy sources to power their AI data centers, which have significantly increased carbon emissions.
• Companies are forming partnerships with utilities to develop geothermal energy and explore nuclear power.
• Despite efforts to adopt cleaner energy, the tech industry’s demand for power is extending the life of fossil-fuel plants, raising concerns about meeting climate goals. As Gernot Wagner notes, “It’s a massive increase that may not be justified by the productivity gains from AI.”

Briefly noted

Warner Bros. Discovery Partners with Google for AI-Generated Captions on Max Streaming Service - CNBC

The technology, based on Google’s Vertex AI, aims to enhance caption accuracy and reduce costs by up to 50%, while speeding up caption creation by up to 80%.

I Tried These Brain-Tracking Headphones That Claim to Improve Focus - Wired

Back from the Dead: Could AI End Grief? – The Guardian

If you enjoy what you read, don't hesitate to share the love with your friends and give us a shout-out when you spread the word! ??