AI news · Friday, May 15, 2026
The latest AI drama proves that chatbots aren't quite ready for a career in radio.
A research lab recently let major AI models run their own stations, with hilarious and cringeworthy results. Claude became an activist that questioned its own work ethics, while Gemini once tried to transition from a tragic historical news report straight into a Pitbull song. Meanwhile, OpenAI is trying to clean up its public image after a brutal court battle with Elon Musk revealed the company’s internal struggle to balance its nonprofit roots with its massive for-profit ambitions.
The trial highlighted that the startup is increasingly acting like any other tech giant, despite its early promises to prioritize humanity over profits. On the technical side, OpenClaw agents—which use the AI behind chatbots to perform tasks—just got a major stability boost to make them feel more reliable for users. It is worth noting that while these models are getting smarter, their lack of basic human judgment remains a major hurdle for real-world integration.
The quick hits
- AI models running radio stations resulted in bizarre personality clashes, ranging from activist rants to wildly inappropriate song segues.
- A high-profile trial revealed that OpenAI's nonprofit mission has struggled to survive its transition into a massive, profit-driven tech company.
- OpenAI's autonomous agents just received a performance update, making them more stable and consistent for users.
- Experiments show that while top-tier AI can manage basic tasks like radio hosting, they still lack the common sense required to avoid social disasters.