Corpus.
Also known as: Corpora (plural)
All of the inputs an AI model has ingested during training and testing. Inputs could be images, data, text, audio, or songs. These can be structured, like labelled rows of data, or unstructured, like a folder of random images. What gets added into a model's corpus has implications for that model's future performance, accuracy, and bias. Garbage in, garbage out starts here and bias too. Not all AI corpora were gathered with permission. Court cases are now emerging where authors, creators, artists, and publishers are bringing claims against major AI labs for copyright infringement arguing their works were taken and ingested into model training without their permission.
In the wild
“Every AI's corpus will be different, because it is humans who decide what kind of data they want to train an AI on.”
Seen here: Fast Company, 2023
For AI users
Ask or at least consider what's actually in a vendor's corpus as that will influence what you get out of the other end. Check out what legal cases are open against the labs whose models you use.
For business leaders
Some AI legal risk sits here. You wouldn't use a supplier who'd stolen ingredients to make the biscuits for your office, so consider the same sense-check for AI and add some sensible steps into your AI procurement checklist including checking what's in a model's corpus. A model provider (lab) should be able to give you details if asked. Bodies like B Lab haven't yet published formal guidance on responsible AI use, but we expect responsible AI badges and checklists to come.