AI Researchers Debate AI's Role and Limitations at First-of-Its-Kind Conference

Imagine a science conference where robots wrote all the research papers!

That's kind of what happened at a new event called Agents4Science 2025.

Scientists let AI, like ChatGPT and Gemini, do the work of leading research and writing up findings.

Some AI helped brainstorm and analyze data, which is pretty cool!

But, the human scientists still had to fix things.

The AI sometimes made up information, like fake references, and didn't always understand the big picture.

One expert felt the AI's research was correct but not very exciting or important.

Another worried people might trust AI too much because it sounds so sure of itself.

Even though AI is getting better, humans are still needed to guide it and make sure the science is actually good and accurate.

The experts think AI could be a great helper, like a super-smart intern, but humans need to stay in charge.