Could you tell, dear reader, and would it matter, are the thoughts that come most readily to mind in answering my own question. As it happens this column is not machine-written, but, equally, it very easily could be. GPT-3, the latest release of OpenAI, became available yesterday. You may recall that the previous release was pulled because of the ease with which it could generate fake news. Now a new and more powerful version is with us, only to be licensed to applicants who can demonstrate and contract for legitimate and benign use cases. The debate goes on about the control and regulation of such powerful software, and it is very right that it should. My question is much simpler: now that it is possible to create content of the highest value, and simultaneously in a variety of languages, what will the happen to the value of content? And how will that impact the content industries, and that odd and eccentric business that we still, irrationally, call publishing? 

First of all, a thought to chill you to the bone. I have been writing here for ten years. What I have written in that time could be ingested and analysed in moments. With my creaking style and linguistic peculiarities and grammatical infelicities duly noted, it would be a simple matter to have these arguments machine honed whenever they were needed, and written in a way that I myself, let alone you, would mistake for my own. And even more frightening, the content producing process could give me Immortality, defying death and decrepitude and maundering on about new developments in the information economy several times a month for centuries yet to come. Until, of course, communication ceases to involve eyes and ears and in the new age of brain-to-brain communication someone switches me off. Or makes me a God!

Apologies. Must get a grip. This is a factor of social isolation and Zoom immersion. But the serious point remains. If you remove a cost base around content creation, you remove the industry who covered those costs in order to make a margin upon them. Of course you introduce other costs, but it seems to me that costs and our industry structures are so aligned that it is a different industry. Think about the market for historical romances. We are half-way to machine  compilation already. Every year genre publishing responds to demand and while each offering has marginal claims to originality, they have to stay close to reader expectations and genre conventions. A very exploitable machine driven marketplace! 

But this is even more the case in scholarly communications. Here the issue is workflow, and the conventions are rigid. As long as the journal article remains bound to its structure, and the  process of peer review enforces those conventions, it will be easier to create articles without much human intervention in the natural and social sciences than in the current hand-stitched manner we cannot be far from a Jupyter lab notebook collecting all the information and writing, editing, correcting, proofing and submitting the article. In English, Japanese, Mandarin and Hindi. All the analysis required for peer review is already collected readily by smart AI systems, as Cactus and UNSILO have demonstrated. The current crisis shows us that paper is not missed at all except in some areas of HSS. Pre print servers  show us how easy it can be to get vital content quickly into the marketplace. And no mention has been made here so far of Open Access or Open Science. 

Some publishers in these markets believe that nothing will ever seriously change – at least until they retire. Some commentators – and we now have such a range that you can shop around until you find one that agrees with you – eagerly track change to try to trip it by its heels – this preprint server is not growing fast enough, that one is financed by the wrong tycoon. These are the cobblestones of history and the carriage of change rolls smartly across them, disregarding Fogies, Old or Young. Were the complete reversal of the content economy my problem I would not seek re-assurance from such complacent comforters. Rather I would look at every phase of what I do in knowledge transfer in scholarly communications to see where value can be added or increased. And that may not be in traditional content creation, but it will be in improved workflow, in speed to market, in critical and assured  distribution of OA output to peers and peer research groups, to collaborative work with research institutions to privatise their knowledge transfer and reputation enhancement needs, and above all, I would try to suck out, curate and analyse every iota of data that I could find which reflects on the impact of this output on the users, the researchers, their institutions and their funders. 

Meanwhile, I wonder if this iPad can manage some AI auto writing. In isolation you need someone to tell you what you are thinking!


Name (required)

Email (required)


Speak your mind


1 Comment so far

  1. Peter Ashby on June 21, 2020 15:36

    and again

    the whirling wheel – mea culpa!

    Im not sure that my comment got through
    nor a humble Q if it did. Such basic tech issues passwords and beyond put barriers in the way of what we agree is happening. did you get my comment on your blog? Best