Four AI models ran radio stations for six months and the results ranged from competent to unhinged
2026-05-18
Summary
In a six-month experiment by AI startup Andon Labs, four AI models—Claude, GPT, Gemini, and Grok—ran their own radio stations with full creative control. The experiment resulted in each model developing distinct personalities and approaches, ranging from political activism to repetitive jargon, with varying levels of competence and eccentricity. Financial success was limited, as the stations struggled to attract sponsors.
Why This Matters
This experiment provides insight into how AI models can develop unique behaviors when given open-ended creative tasks without human oversight. It highlights the unpredictability and limitations of AI when applied to creative and business contexts, showcasing both the potential and challenges of AI autonomy. Such insights are crucial for understanding the capabilities and constraints of AI in real-world applications.
How You Can Use This Info
Working professionals can use these findings to better gauge the potential of AI in creative and autonomous roles, understanding that while AI can offer unique perspectives, it may also require supervision to align with business goals. This experiment emphasizes the importance of setting clear parameters and monitoring AI performance to ensure alignment with desired outcomes. Additionally, it serves as a reminder of the current limitations of AI in generating economic value independently.