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

Arvind Narayanan, a computer science professor at Princeton University, is best known for calling out the hype surrounding artificial intelligence in his Substack, AI Snake Oil, written with PhD candidate Sayash Kapoor. The two authors recently released a book based on their popular newsletter about AI’s shortcomings.

But don’t get it twisted—they aren’t against using new technology. “It’s easy to misconstrue our message as saying that all of AI is harmful or dubious,” Narayanan says. He makes clear, during a conversation with WIRED, that his rebuke is not aimed at the software per say, but rather the culprits who continue to spread misleading claims about artificial intelligence.

In AI Snake Oil, those guilty of perpetuating the current hype cycle are divided into three core groups: the companies selling AI, researchers studying AI, and journalists covering AI.

Hype Super-Spreaders

Companies claiming to predict the future using algorithms are positioned as potentially the most fraudulent. “When predictive AI systems are deployed, the first people they harm are often minorities and those already in poverty,” Narayanan and Kapoor write in the book. For example, an algorithm previously used in the Netherlands by a local government to predict who may commit welfare fraud wrongly targeted women and immigrants who didn’t speak Dutch.

The authors turn a skeptical eye as well toward companies mainly focused on existential risks, like artificial general intelligence, the concept of a super-powerful algorithm better than humans at performing labor. Though, they don’t scoff at the idea of AGI. “When I decided to become a computer scientist, the ability to contribute to AGI was a big part of my own identity and motivation,” says Narayanan. The misalignment comes from companies prioritizing long-term risk factors above the impact AI tools have on people right now, a common refrain I’ve heard from researchers.

Much of the hype and misunderstandings can also be blamed on shoddy, non-reproducible research, the authors claim. “We found that in a large number of fields, the issue of data leakage leads to overoptimistic claims about how well AI works,” says Kapoor. Data leakage is essentially when AI is tested using part of the model’s training data—similar to handing out the answers to students before conducting an exam.

While academics are portrayed in AI Snake Oil as making “textbook errors,” journalists are more maliciously motivated and knowingly in the wrong, according to the Princeton researchers: “Many articles are just reworded press releases laundered as news.” Reporters who sidestep honest reporting in favor of maintaining their relationships with big tech companies and protecting their access to the companies’ executives are noted as especially toxic.

I think the criticisms about access journalism are fair. In retrospect, I could have asked tougher or more savvy questions during some interviews with the stakeholders at the most important companies in AI. But the authors might be oversimplifying the matter here. The fact that big AI companies let me in the door doesn’t prevent me from writing skeptical articles about their technology, or working on investigative pieces I know will piss them off. (Yes, even if they make business deals, like OpenAI did, with the parent company of WIRED.)

And sensational news stories can be misleading about AI’s true capabilities. Narayanan and Kapoor highlight New York Times columnist Kevin Roose’s 2023 chatbot transcript interacting with Microsoft’s tool headlined “Bing’s A.I. Chat: ‘I Want to Be Alive. 😈’” as an example of journalists sowing public confusion about sentient algorithms. “Roose was one of the people who wrote these articles,” says Kapoor. “But I think when you see headline after headline that’s talking about chatbots wanting to come to life, it can be pretty impactful on the public psyche.” Kapoor mentions the ELIZA chatbot from the 1960s, whose users quickly anthropomorphized a crude AI tool, as a prime example of the lasting urge to project human qualities onto mere algorithms.