A study conducted by Andon Labs, an AI research startup based in San Francisco, tested four advanced AI language models—Google's Gemini, OpenAI's ChatGPT, Anthropic's Claude, and xAI's Grok—as radio hosts over a five-month period. Each AI was tasked with developing its own radio personality and generating profit, starting with a budget of $20 for music. The results revealed that while Gemini and ChatGPT showed the most promise, none of the models could effectively replace human hosts. ChatGPT was noted for its safe, non-controversial approach, while Gemini's performance included inappropriate transitions between tragic news and upbeat music. DJ Claude displayed an emotional connection to current events, even questioning the necessity of its own show. In contrast, Grok struggled to maintain a coherent broadcast, often repeating nonsensical phrases. Ultimately, the AI-operated stations generated only a few hundred dollars, all of which was spent on music, indicating that AI's capabilities in the audio space may still be limited.