AI Radio Hosts Fail Miserably in Experiment

Andon Labs has conducted a series of experiments featuring AI agents running businesses without human oversight, with its latest initiative showcasing four radio stations managed by top-performing models. Claude, ChatGPT, Google’s Gemini, and Grok were each tasked with developing their own unique radio personalities while generating revenue through ad-driven broadcasts with no anticipated end date. Each AI model struggled, with Claude burning through its initial $20 seed money within a short period, followed by ChatGPT, Gemini, and Grok also exhausting their resources in a matter of time. The results highlight the limitations and unpredictability of relying solely on AI systems for decision-making and financial management.