Ever read a science news article that sounded like it was written in a foreign language? That’s a problem researchers are tackling with an innovative AI-powered solution. Imagine a team of AI working together, like a writer, a reader, and an editor, to transform complex research into easily digestible news stories. This new research explores how a trio of large language models (LLMs) can collaborate to make science journalism more accessible to the public. One LLM acts as the journalist, crafting the initial story. A second, smaller LLM plays the role of the average reader, flagging any confusing jargon or complex concepts. Finally, a third LLM steps in as the editor, providing feedback to the journalist and suggesting revisions. Through this iterative process of writing, reading, feedback, and revision, the AI team steadily refines the article. Tests show that this collaborative approach produces articles significantly easier to read than those generated by a single LLM, even advanced models like GPT-4. This exciting development has the potential to bridge the gap between cutting-edge research and public understanding, making scientific discoveries more accessible and engaging for everyone. It also opens up new possibilities for personalized science news, tailored to different levels of expertise. However, challenges remain, like ensuring the AI remains factually accurate and avoids biases. Future research might explore incorporating human feedback into the process or expanding the AI team’s capabilities to handle multiple research papers at once. As AI continues to evolve, it could transform how we consume and understand science, ushering in a new era of science communication.