Claude in Science: Anthropic Expands AI for Science

The Space Telescope Science Institute, Wellcome Sanger Institute and Francis Crick Institute are also amongst the described organisations. These examples demonstrate the breadth of the programme, but the source provides no general application guarantee or individual credit amount for each researcher.

What Researchers Do with Claude

According to Anthropic, scientists use Claude to formulate hypotheses, design experiments, analyse complex data and communicate findings. The company reports supporting projects in biology, medicine, physics, chemistry, climate and mathematics, amongst others. These are reported use cases, not an independent comparison of research quality ^[1].

A concrete example is the Model Hardware Standard, developed with HHMI Janelia Research Campus. This shared specification is intended for AI agents and laboratory instruments. Anthropic explicitly states that MHS is still in research preview. According to the report, the Allen Institute is working on agent teams for analysing open research data ^[1].

Open Resources and Human Responsibility

Anthropic expects many projects to yield public resources, such as datasets, open-source software and shared standards. That is a goal and expectation; not every mentioned result is thereby already available.

Our editorial recommendation: assess access and costs separately from scientific reliability. Free credits do not replace oversight of data, methods or reproducibility. Document which conclusions come from the researchers themselves and which parts have been proposed by a model. This keeps the scientist responsible for the interpretation.

Read More

Also read Claude in Drug Research.