Taylor & Francis

Taylor & Francis, the British academic publisher whose roots reach back to 1798 and whose imprints include Routledge and CRC Press, publishes thousands of scholarly journals and books a year. Its parent company Informa disclosed in May 2024 a roughly $10 million agreement giving Microsoft non-exclusive access to its “advanced learning content and data” to improve AI systems — one of the first big AI licensing deals in academic publishing.

Sold without asking

The trouble was who wasn’t asked. Academic authors learned of the deal from the news months later, and reported they were given no notice, no opt-out, and no additional payment. Leiden University researcher Ruth Alison Clemens warned the bulk data sale set “a new and worrying precedent”; Georgia Tech’s Lauren Barbeau was shocked that “no one had said anything to us at all.” The Society of Authors objected to publishers signing tech deals without consulting creators, and some academics moved to pull their work from Taylor & Francis journals. The publisher countered that the partnership would help develop tools to “improve research” and that authors’ royalty arrangements covered such subsidiary uses.

The deal became the reference case for a wider wave: Wiley, Sage, and others soon disclosed their own AI licensing arrangements, and Informa deepened its Microsoft partnership in subsequent years. Defenders — including open-access advocates like SPARC’s Heather Joseph, who called training AI on high-quality scholarship “an unqualified good thing” — argue the alternative is models trained on worse information.

Taylor & Francis thus sits oddly on the pro side of the ledger: enthusiastic about AI’s value for research and for its balance sheet, while embodying the consent problem that fuels the other side of the debate.