From the press release announcing Clarivate’s partnership with AI21 Labs (“a pioneer in generative artificial intelligence”):

The collaboration will integrate large language models into solutions from Clarivate, to enable intuitive academic conversational search and discovery, specifically designed to foster researcher excellence and drive success for researchers and students, while adhering to core academic principles and values. AI has the potential to revolutionize the world, but its effectiveness relies heavily on the quality of the training data. With billions of trusted, curated, articles, books, documents and propriety best in class data points, Clarivate is well-placed to lead the market on this opportunity, providing customers with the highest quality open, licensed and proprietary content, data and insights while mitigating associated risks.

So here’s another surveillance business to layer on top of all the Web of Science/ProQuest subscription revenue: selling scholars’ behavioral data as AI training data. Insult to injury.

Simmons University, a Boston women’s college, is cutting literature, philosophy, and sociology, among other majors—according to Inside Higher Ed. Enrollment is way down, but another challenge is the institution’s for-profit online-ed “partner”:

President Lynn Perry Wooten is also trying to change a contract Simmons has with 2U, an online learning company. Under the contract, negotiated before Wooten arrived at Simmons, 2U gets between 50 percent and 62 percent of tuition for each student in its online programs—more than half of all Simmons’s graduate students. The contract runs through 2039.

Steven Inchcoombe, “Chief Publishing Officer” at Springer Nature, in a January Scholarly Kitchen interview, asked what publishing innovation he’s most proud of:

For me, I look at the work we have been doing using AI to enable summarizations (for different levels of knowledge), language improvement and auto-translation, structured support of peer-reviewers, helping book authors survey the literature, and addressing integrity concerns such as software to spot plagiarism, tortured phrases, and image manipulation. Clearly humans still play an important role in these use cases but technology is enabling us to be more adaptive and to operate at a scale that we could previously only dream of.

Machine learning as labor-reducing and margin-fattening “innovation,” or: Why 35 percent profits aren’t high enough.

From the Open Access Week announcement:

“Community over Commercialization” is the theme for this year’s International Open Access Week (October 23-29). This theme encourages a candid conversation about which approaches to open scholarship prioritize the best interests of the public and the academic community—and which do not. […] Adopted by its 193 Member States, the UNESCO Recommendation on Open Science highlights the need to prioritize community over commercialization in its calls for the prevention of “inequitable extraction of profit from publicly funded scientific activities” and support for “non-commercial publishing models and collaborative publishing models with no article processing charges.”

The huge diamond OA gathering in Mexico is, of course, the same week—with Europe coming to Latin America. In more ways than one, as it turns out: Once the world’s center of APC activity, Europe is now (or so it seems) charting an open-authorship future. And it’s explicitly calling for nonprofit infrastructure, here again coming home to Latin America. Also in gestation: cOAlition S pulling the plug on transformative [sic] deals, and U.S. Nelson Memo implementation.

I’m heating up the popcorn.

The Ivy Plus Libraries have come out swinging against APCs in their Nelson Memo statement:

It is in that spirit that we want to highlight the dangers of allowing the interests of commercial publishers to dictate the paths available to implementing this bold new guidance on open scholarship. We refer here to the pay-to-publish model of open access to research publications, as exemplified by individual APC (article processing charge) fees charged directly to authors, and/or institutional Read and Publish agreements where libraries pay bulk APCs on behalf of their scholars and unlock institutional access to read pay-walled content. […] Implementing the Nelson memo via an APC model is antithetical to the equity goals so clearly articulated in the guidance memo and the values of our institutions.

The institutions—which include MIT, Stanford, Chicago, and Duke, among others—cite the exclusion of authors:

Locking in a norm where an author, funder, and/or institution must pay an opaque and often costly fee for the right to publish an article risks locking out scholars from less resourced institutions and less well funded disciplines. The equity issue in the APC model extends globally for authors and researchers in lower-income countries who must navigate publishers’ convoluted and demeaning APC waiver procedures that may result in denial of the waiver or discounted APC fees that are still unaffordable. Equitable opportunity to contribute to scholarly literature is as important for the integrity and usefulness of scholarship globally as is the open accessibility to read.

The open-authorship movement continues to grow.

Alex McLean, writing for on his blog about the treasure hunt required to find OA copies of his new book, despite a $15,000 subvention to MIT Press:

Although all copies of the book are openly licensed into the creative commons, some are nonetheless paywalled. Indeed, the ebook is has wide digitally distribution to the kindle, apple, google, kobo etc ebook stores where you can buy access to this creative commons license for $25.99. Unfortunately, the digital rights management (DRM) imposed by the store makes it difficult to benefit from the freedom to share and modify the text that the open license grants you. So really, you are paying $25.99 to lose benefits.

And:

Worse, the MIT Press website steers you towards these digitally-rights-managed, $25.99 paywalls and away from the otherwise identical free-to-download ebook that we paid the subvention for. If you click the big ‘ebook’ button, which tantilisingly has no price next to it (screen shot below), you are directed to the Penguin Random House commercial distribution of it:

To track down a free ebook,

you have to instead click on the ‘resources’ tab, and find a link to the epub or mobi ebook download there. Of course, this isn’t a mere resource for the book, but the actual book, so that’s a bit like hiding the free download behind a door that says ‘beware of the leopard’.

The PDF version is on still another page, where it’s divided into 15 separate chapter downloads.

McLean considers this buried-OA a “kind of performance done to placate” funders—a fair read. MIT, which is rightly praised for its fee-free OA initiatives, cites the need to recoup costs and promises some minor fixes to their standard layout. Good. McLean’s point still holds: the full PDF (single download) and ebook versions should be at least as prominent as the paid options.