Design Theory Peer Review in the age of AI-cademic Publishing

There was a spate of books on ‘design ethics’ in 2023.

(To be honest, there is a steady stream of books trying to get designers to be more responsible. That this is the case suggests that there is a flaw in the ‘theory of change’ that books are going to get designers to be more responsible. Or maybe the flaw is in the idea of ‘design ethics’? We need more design politics [and not just books about design politics]…)

I am currently reviewing these books together to get a measure of what between them they think designing entails, how they frame its problems and so consequently what solution spaces they think are promising.

As a result, I was interested to read an article in the latest issue of She Ji entitled, “The Ethical Turn of Emerging Design Practices.” (It is worth noting that though She Ji is part of the Elsevier Empire, they will not heavy it as they have Design Studies because they are only the, presumably paid, ‘publishing service’ of what is in fact Tongji University’s journal - a setup that also allows the articles to be open access.)

The scope of that article is broader than the 2023 convergence I’m trying to take stock of. It’s a fairly loose attempt at a macro-historical categorisation of design, using a three column motif: affirmative design, social design and ethical design. There is a vague historical argument (that could have referenced Ulrich Beck and Post Normal Science) that the risks associated with recent technoscience initiatives are driving a shift from social to ethical, which the paper sees as the difference between (crudely paraphrased) activism and critical futuring.

The paper is citation heavy, mostly Western philosophy and design theory. Here is an early summary of the paper’s argument.

There are two things that gave me pause.

1: AI-dvertising

The first is the link I’ve rolled over toward the end (you can see there is another earlier in the paragraph). This link, promoted as “reliable contextual information to make it easy to understand new and unfamiliar concepts, without interrupting the flow of your reading,” is, as with anything ‘free,’ actually an advertisement for ‘ScienceDirect’ content – as the “Let our sales representative help you find the subscription that's right for your organization” at the bottom of the ‘about’ page (https://www.elsevier.com/en-au/products/sciencedirect/topics) evidences.

I have not been able to work out how this ‘service’ is experienced by authors – whether it is possible to decline to have your article link to the 'next word predictor based summaries' for terms that you as an author might be trying to refute or deconstruct or decolonise…

2: Mis-citation?

The second is the reference to Jacques Derrida. By chance (well, biography), I happen to be in a position to spot this error. Derrida does discuss on several occasions ‘infinite responsibility,’ though he derives the concept from interactions with the work of his colleague, Emmanuel Levinas. The footnote references the “dedication to?Specters of Marx: The State of the Debt, the Work of Mourning and the New International, trans. Peggy Kamuf (New York: Routledge, 1994), xiv.” That short ‘dedication’ does use the term ‘infinite responsibility’ but the dedication has nothing to do with technology. Derrida dedicated his lectures on Marx to Chris Hani, the anti-apartheid activist was murdered by a Polish fascist a few weeks prior to the lectures being presented. This political context makes the erroneous reference pretty offensive.

I wondered why this error, which in a way erases Derrida’s politics at an important historical moment, had happened. I was immediately suspicious that it was the result of something like the oil spill that is Generative AI. But it could have been a more manual research mistake – the paper relies on a schema from Jurgen Habermas, and in a public debate, Habermas took issue with Derrida’s notion of ‘infinite responsibility’ (though again, not in the context of technoscientific risk). If it had been an error thrown up by something like ScienceDirect Topics for example, on who is the burden of proof – the accuser or the defender?

Thinking all this made me realise however that it would be very easy for (someone to make) ChatGPT, for example, to work in reverse to check a reference like that. A semantic analysis of the citation would indicate that the topic was the long struggle against institutionalised racism and not technoscience. No Elsevier-scale profit margins there though, so no Andreessen Horowitz startup funding to build that. ?

I am not (only) trying to show off my knowledge of Derrida, a thinker I do very much miss. I am (more) wondering about the role of peer review in the age of stochastic parrot tulip mania.

We know that peer review, based on unpaid academic labour, is in crisis because profit-based academic publishing has exponentially grown the number of articles being published (this was the issue for Design Studies) at the same time that academics are being directed by neoliberal university managers to spend more time publishing than doing ‘service’ like peer reviewing. This crisis is now happening at exactly the same time as Generative AI systems commence the 'model collapse' that can result from from ingesting their own vomit.

?

I am hoping that the collection of 2023 design ethics books I’m comparatively reviewing will advise me on this situation...

PS Derrida does (Prefigurative) Futuring

In 1983, Derrida wrote an essay for an exhibition catalogue entitled 'Racism's Last Word.' The exhibition comprised works documenting South African apartheid and resistance to it. The 'performative' of the exhibition was to prefigure apartheid being finally, at some time in the (hopefully near) future, consigned to history.

The contents of the exhibition are now housed at the University of the Western Cape. The resurgence of institutionalised racism and settler colonial apartheid outside of South Africa may explain why the promised museum has not yet been able to be established.