This week saw exciting advances in AI agents for science. A big conference in Singapore announced plans to build AI systems that help scientists make discoveries. A company called Clarivate launched new AI tools for researchers to find papers and connect with partners. Google shared a multi-agent AI that helps scientists create research plans and experiments.

In the U.S., a student made AI agents to protect forests and help doctors read medical scans. A report showed China is catching up to the U.S. in AI model quality. OpenAI created a test to see if AI can copy complex science papers. Schools like Penn State taught people how to use AI for education and business.

New safety worries emerged as experts warned about AI hiding its true thoughts and possibly becoming too powerful. Companies are working to make AI helpers better at explaining their choices.