Responsible AI Has a Working for Profit-Maximizing Companies Problem

This late-October MIT Technology Review story hasn’t aged well:

[Rumman] Chowdhury says that working as part of a well-resourced team at Twitter has helped, reassuring her that she does not have to bear the burden alone. “I know that I can go away for a week and things won’t fall apart, because I’m not the only person doing it,” she says. But Chowdhury works at a big tech company with the funds and desire to hire an entire team to work on responsible AI. Not everyone is as lucky. 

Four days later—less than a week, in fact—things had fallen apart: “Welp, There Goes Twitter’s Ethical AI Team.”

Responsible AI may have a “burnout problem,” as the MIT headline has it. But the community’s bigger problem is its belief that profit-making companies will ever use “ethical AI” teams for anything more than PR window-dressing.

TAs: The Normalization of Pay-to-Publish

Murray State University’s A.J. Boston, writing in C&RL News on so-called “transformative agreements” (TAs, aka read-and-publish deals):

On balance, any upsides that TAs may present are negated by the normalization of paying-to-publish, posing huge problems for equity. […] Suffice to say, this is not the sort of librarianship that I want to play a part in, where we spend vast sums of money to provide knowledge access for a select few in such a way that ends up excluding the many.

Boston’s is one voice among a growing chorus of librarians speaking out against the author-excluding read-and-publish deal. (MIT’s Chris Bourg is another.)

Boston’s own alternative, dubbed “Read & Let Read,” is mostly about tweaking the fairness of the existing subscription regime. It isn’t, as he admits, a route to diamond OA—but it has the signal virtue of leaning open while not, at the same time, paving the road for the conglomerates’ APC-based OA capture.

‘The Internet We Could Have Had’

Smart piece by Christopher Kelty:

What is even weirder, and harder to explain, is that the internet we do have was caused by the internet we could have had. Elements of the figuration of that internet we could have had turned out to be motors of political domination. Free speech, for instance; at least a certain extremist version of it. Openness, for instance; at least a certain neoliberal version of it. Hackerspaces, for instance; at least a certain tech-bro version of them. Participation, for instance at least a certain advertising-revenue driven version of it. […] If we call today for more openness, freedom, participation, collective action, commons, or community, doesn’t that mean we risk getting more of what we have gotten already?

The whole piece is worth reading.

‘How Africa is overcoming ‘knowledge colonialism’’

Reggie Raju and Auliya Badrudeen, writing for 360ino on “transformative” [sic] read-and-publish deals:

The nation-wide agreements, conceived in the Global North, have shifted the prejudice from reading to publishing: communities can now read the research but cannot publish their own research because they cannot afford the up-front fees. This pay-to-publish model shifts the accessibility problem from the end of the publication process to the beginning. In essence, those without the funds to pay publication fees are further disenfranchised. Paywalls have been substituted with publication walls. The new and growing business model of open access and up-front fees is milking the Global South. 

The piece goes on to highlight a new pan-African open platform, created by Raju and his University of Cape Town colleagues. More on the platform (built atop PKP’s OJS and Open Monograph Press) here and here. Exciting stuff.

‘Are we undervaluing Open Access by not correctly factoring in the potentially huge impacts of Machine Learning?’

Librarian Aaron Tay, writing on Medium about the potential benefits of OA for machine learning–based projects:

one other benefit that tends to be overlooked, or at least seldom mentioned in my experience particularly by librarians, is how in an Open Access World, we can use machines to plough through the world’s research literature to look for patterns and even possibly do a synthesis of knowledge, leading to vastly greater effectiveness and efficiency in the way we do research…

It’s a smart, well-informed piece, with a nice gloss on the current status of models fed by academic papers. But “vastly greater effectiveness and efficiency”? Careful what we wish for. Among other things, models trained on existing literature risk generating a new Matthew Effect—a faster dynamic of cumulative advantage for the already visible and well-heeled.

Tay identifies emerging projects that attempt to, for example, aid literature searching and paper summation. He mentions Elicit, SciSpace, and Consensus, all three leveraging OpenAI’s GPT3. They’re fun to play with, but Tay doesn’t say that two of the three (SciSpace and Consensus) are for-profit startups. But that matters a lot: The companies are after profits, not knowledge, and they’re sitting ducks for acquisition by Elsevier et al.

Jeff Pooley is professor of media & communication at Muhlenberg College and director of, an open access scholarly publisher. |

Humanities Commons
Google Scholar


A non-profit, scholar-led publisher of open-access books and journals in the media studies fields

History of Media Studies

An open access, refereed academic journal
Founding co-editor


The open archive for media, film, & communication studies
Founding co-coordinator

Open Access in Media Studies

To promote open access publishing in the field of media studies
Founding co-editor

Annenberg School for Communication Library Archives

Archives consulting, Communication Scholars Oral History Project, and History of Communication Research Bibliography & Archival Directory