To help an artificial intelligence (AI) algorithm develop legal skills, Harvard Law School announced that it has made over six million court cases available to the general public.

The project, dubbed the Caselaw Access Project (CAP), included the publishing and digitizing of every reported U.S. state and federal legal case dating back from the 1600s all the way through 2017.

The project, which took over five years to complete, covers roughly 6.5 million individual cases and is expected to be a significant resource for audiences such as researchers, the general public, the legal community and developers of legal AI applications.

To train their software, companies that build legal AI software require data. However, in the past, developers had to construct their own databases, using the limited amount of data available to the public.

The data from Harvard provides these companies with a resource to train their AI instead. Users of the Caselaw website will be able to browse and download specific case files or bulk files in zip folders.

Experts predict that the resource will be an invaluable tool for short-staffed legal aid companies, granting them access to considerable information for free.