DAVE LANKES: Is AI Making Us Work More?

The promise of AI has always carried an implicit bargain: work smarter, not harder. A recent UC Berkeley study by Aruna Ranganathan and PhD student Jing-Kee Magi-Yay suggests the reality may be quite different. Their eight-month study of a 200-employee company found that far from lightening the load, AI is intensifying work in ways that are hard to see coming.

Workers in the study operated at a faster pace, took on a broader scope of tasks, and put in more hours—often voluntarily and without being asked. This isn’t the AI-assisted leisure future that was advertised. Instead, the efficiency gains seem to be absorbed immediately by an expanded definition of what a worker is expected to accomplish.

The findings appeared in Harvard Business Review and should give pause to anyone who thinks AI adoption is automatically a quality-of-life win for employees. As the study’s authors report, the pace and scope of work both expanded alongside AI use. Think of Matthew Broderick taking a day off in that Super Bowl ad—a charming fantasy that this research suggests is more myth than forecast. The real question for organizations isn’t just “what can AI do?” but “what should AI be allowed to do to our people?”

Ranganathan, A., & Ye, X. M. (2026, February 9). AI doesn't reduce work—It intensifies it. Harvard Business Review.

BETH PATIN: ChatGPT Gets a Sponsor

We probably all sensed it was coming. ChatGPT is now testing advertisements, starting at the free tier and lowest subscription levels—and the implications for how we think about AI-generated answers deserve serious attention.

This Super Bowl season offered a sharp preview of what’s at stake. An Anthropic commercial—which reportedly irritated OpenAI’s Sam Altman—depicted a man asking a ChatGPT-style assistant for advice on repairing his relationship with his mother. The AI’s warm, therapist-like guidance pivoted quickly into a pitch for a cougar dating app. It was satirical, but the underlying critique landed: when ad revenue enters the picture, the line between helpful guidance and commercial nudge gets dangerously blurry.

The numbers are significant. OpenAI is reportedly charging three times Meta’s rate per thousand impressions, with a minimum $200,000 buy-in—ad revenue on par with Netflix. For now, the company promises ads won’t influence answers. But as librarians know better than anyone, the source always shapes the message. If you’re recommending ChatGPT to patrons as a place to get reliable information, it’s time to factor in what’s now funding the answers.

OpenAI. (2026, February 9). Testing ads in ChatGPT. OpenAI.

Perez, S. (2026, February 9). ChatGPT rolls out ads. TechCrunch.

MIKE EISENBERG: AI on the Case — and in the Cuffs

In keeping with this episode’s theme of crime and information, I put several AI systems to work on a question: how is AI actually being used in crime solving? The responses were detailed, informative, and—refreshingly—careful.

AI contributes meaningfully across several investigative fronts: digital forensics, pattern detection, cold case analysis, fraud detection, timeline reconstruction, and deepfake validation. It can surface deleted files, link unrelated incidents, and match faces to databases. The AI tools I queried were also candid about the limits: accuracy degrades with poor lighting and camera quality, and demographic bias in facial recognition systems is a documented problem.

The harm cases are real and recent. Jason Killinger was held for eleven hours in Nevada after a casino’s facial recognition system misidentified him—despite valid ID. Travis Williams was wrongfully arrested by the NYPD in 2025 under similar circumstances. Robert Williams’ 2020 Detroit case remains a landmark example of the technology’s failures. The AI systems I consulted put it plainly: AI is a super-fast analyst, not a super-intelligent detective. It surfaces patterns; humans must still interpret them. That distinction matters—especially when the stakes are someone’s freedom.

Biscobing, D. (2026, February 11). Man ‘falsely arrested’ for cold case with facial recognition threatens $3 million lawsuit. ABC15 Arizona.

Morioka, S. (Winter 2024-2025). Flawed facial recognition technology leads to wrongful arrest and historic settlement. Law Quadrangle.

Cramer, M., & Hill, K. (2025, August 26). How the N.Y.P.D.'s facial recognition tool landed the wrong man in jail. The New York Times.