AI is eating the academic publishing industry alive, but some good might come of it

It’s a challenge for researchers to explain academic publishing to someone outside of a university. Whoever they’re speaking to has to work very hard to remain polite while concluding that the researcher is either naïve or very stupid. The whole system sounds very much like a scam.

It goes like this: academics, employed by universities, conduct research. They write up their findings. Then they hand over their work, and often the copyright for that work, to commercial publishers – for free. Those publishers send the work out to other academics to review, also without payment. There’s usually a bit of back-and-forth as the researchers and their peer-reviewers decide whether or not the article makes a quality contribution to the field.

The publishers then sell the resulting articles at extortionate prices to anyone keen to access cutting-edge knowledge – including to the university that paid for it all.

If the researcher wants their work to be available to the public without them having to fork out cash, the researcher can opt for “Gold Open Access” by paying the publisher up to thousands of dollars per paper. You read that right. Researchers generally have to choose: either they must pay to share their work widely or their readers have to pay to read it. Whichever choice is made, the researcher, and their employers, will not see a red cent.

And yet they are lining up to get these publications. Begging prestigious journals to take their money in exchange for publishing their work.

At this point, the person to whom the researcher is explaining the system is probably starting to step nervously away.

But it gets worse. In South Africa, public universities are awarded subsidies from the taxpayer’s purse for every publication. So, taxpayer money is going directly to the academic publishing industry. For the university to get these subsidies, the article has to appear in a journal on one of the “accredited lists” and these lists prioritise the Big Five.

The academic publishing industry generates more than $19-billion in revenue annually, more than half of which is in the hands of the Big Five academic publishers – Elsevier, Springer, Wiley, Taylor & Francis, and SAGE. These companies post profit margins in excess of 35%, with some estimates putting margins above 40%. This is a bigger profit margin than Google or Amazon; unsurprising when you consider that most of their overheads – the research labour, equipment and quality control – are all done for free.

But this story is about more than corporate greed. It is about what universities have allowed themselves to become.

Publish or perish

The academic publishing industry built its stranglehold on knowledge dissemination with a great deal of support from the higher education sector. Universities have embraced performance management systems that reduce scholarly life to a metric: publications, citations, impact factors, h-indices. Promotion, retention, grant funding and institutional rankings all flow through the narrow channel of peer-reviewed journal publications, and especially through publications in journals that the Big Five happen to control.

This is the logic of today’s numbers-obsessed university that cannot tell the difference between the creation of knowledge and the production of outputs. A university that cannot tell the difference between the creation of knowledge and the production of outputs has confused use value with exchange value, that is, what knowledge does with what it sells for.

When a university judges its academics only by how many papers they publish, it is feeding the parasitic publications industry. Some universities now have performance reviews that count journal articles regardless of content. And they hand out professorships on the basis of having a set number of articles on their CV. All of this is higher education feeding an industry that has captured the most basic function of the university: the creation and sharing of knowledge. The commodification of knowledge by the publishing industry is entirely entwined with the higher education system itself.

In walks the disruptor

Generative AI has now infiltrated this perfectly constructed neoliberal machine. A major study by Topaz and colleagues, published in The Lancet in May 2026, analysed nearly 2.5 million papers and 97 million citations. They found that in 2023, roughly one in every 2,828 papers contained fabricated references that did not correspond to real publications. By 2025, that figure had reached one in 458. In the first weeks of 2026, the rate was one in 277. Springer Nature retracted an entire book in July 2025 after discovering that it referenced fictitious works.

These are citations to papers that do not exist; plausible-sounding titles, credible-seeming authors, convincing journal names, made up by large language models such as ChatGPT or Claude. Such fabricated references copy the format of genuine citations and include the names of established researchers in the field, along with plausible article titles.

There’s a simple logic to all this. In a publish-or-perish environment where numbers count much more than quality and where impact factors count more than actual impact, there is every reason an academic would hand intellectual engagement and knowledge dissemination over to a large language model.

The elaborate ritual of article submission, review, revision and publication has become so disconnected from the circulation of ideas that academic articles are produced to be counted, not to be read. They are processed by journals for profit, not for knowledge building. And now many of them are being written, in part or in whole, by machines that do not understand what is being said.

Good riddance! (with caveats)

If the exposure of AI-generated slop, hallucinated references, fabricated data and voiceless prose brings about the downfall of the Big Five publishing oligopoly, I cannot mourn it. An industry that profits so spectacularly from unpaid labour, that sells publicly funded research to the public at an enormous mark-up, that has used its stranglehold on prestige to distort what universities are for, well, this industry deserves disruption.

The question is what comes after. If the AI-induced crisis in academic publishing leads universities to rebuild their knowledge-sharing infrastructure around principles of openness, equity and genuine intellectual exchange, then something of value might emerge from the wreckage. If this is the turn that prioritises university and NPO-led Diamond Open Access – free to publish and free to read – and forces a reckoning with the publish-or-perish logic that made the publishing industry so powerful, then we’d all be better off. DM