OpenAI and Apple are negotiating content licensing deals²¹ with news publishers to train their AI systems. But according to sources the deal sizes are smaller than the industry expects. Some publishers feel they aren't getting fair value.

What’s going on here?

OpenAI deal size for access to publishers’ content is smaller than expected.

What does that mean?

We’re talking about offers as low as $1-5 million annually to some publishers. It’s like getting the assistant a coffee to peek inside the boss’s laptop. Apple is saying the “dinner’s on me” but wants to copy stuff to its hard drive. Basically more flexibility in how it uses the content. Google feels behind in the race (again) but let’s not forget that Google already has TeamViewer access with its products like Google News.

Although OpenAI has made some big-name deals earlier (Alex Springer and Associated Press), more publishers will want to hold out and see how the NYT’s case against OpenAI turns out.

Why should I care?

OpenAI (and other AI companies) will push harder to secure these contracts with publishers to a) get their hands on unique data and b) signal their “friendship” with creators. At the same time, it’s likely that more publisher will sue AI companies over the use of their materials.

I wonder if there’s an opportunity for smaller media houses here to go with the AI companies, instead of against them.