Pl@ntNet: the community corrects the machine

You photograph a plant and the app suggests the most likely species. What interests us in Pl@ntNet is the loop that connects the tool to the people who use it.

A tool that learns from its users

Every shared observation can be reviewed by other users, botanists among them. Validated observations are added to the data the recognition model is trained on. The more active the community, the more reliable the tool, and a reliable tool draws more people in.

Here, AI makes botanical knowledge available to people who would never have opened a field guide, and it sends researchers a volume of observations that no field survey could produce.

What it shows

It has what MAUB looks for in a project. The need comes before the technology: knowing what grows around you. The project is run by public research, and its business model does not depend on capturing attention. And the data it produces serves others besides the app itself.

Do you know a project that works this way? Suggest it.