DFace News🧪 OpenAI tests AI agents on real lab tasks
If you work in a biology lab or study wet-lab science, you might soon see AI helping with daily experiments. OpenAI has released BioBasicsBench, a new benchmark that measures how well AI agents handle 100 common lab tasks, from interpreting raw experimental data to routine wet-lab procedures. The benchmark is designed to evaluate whether AI can reliably assist with the kind of hands-on science tha…
Source: OpenAI