Reading Time: 5 minutes
The Eliza psychotherapist emulator  – an updated version of the 60s illusion of intelligence

A welcome return to the NEC (formerly Rave) Conference in London. There was a smaller attendance than in previous years, but as before, the quality of presentations was high, and the presence of the organizers was very discreet – there was (almost) no direct pitching.

As usual, David Smith from the IET gave the keynote, and as usual, his presentation was full of fascinating links that tempted you to start browsing immediately. As for theme, his keynote seemed to be a mixture of exciting and alarming the attendees with the possibilities of AI. The usual mythmaking emerged, for example the Google staffer, Blake Lemoine, who resigned claiming that LaMDA was truly sentient. Was this truly disturbing or was he simply bamboozled by the appearance of cognition? The myth of objects coming to life seems to be embedded in our fantasies, from Pygmalion to Eliza, the pioneering NLP tool that seemed to be knowledgeable simply because it answered your statements with a question. You can see a recent emulation of Eliza here, and it still makes an effect on those seeing it for the first time.

Yet software that creates new images, based on pattern matching, is not so distant from Adobe Photoshop, which has had the ability to paste out or add in people and objects for many years. My feeling is that David Smith’s presentation, which combined the mystical and the threatening with the practical, did not advance the intelligent adoption of narrow AI in academic publishing. We need more case studies of what works in practice, and fewer scary stories.

Matthew Hayes of SAGE Publishing described a useful tool that enhances discovery for libraries. By increasing publisher and library usage, such tools may simply move the existing traffic around a bit – people don’t actually search more than they did before. Still, it demonstrates to the libraries that their traffic has increased.

Mark Loftus, a psychologist from Jyre, apologised for being a late substitute presenter, but had no need to apologise. In a brief presentation about innovation, he revealed some surprising ideas. As humans, we are built to innovate, and yet large organisations have all kinds of procedures in place to prevent innovation. So if you want an organisation to innovate, you have to set up a culture to enable it – and that means more than just asking everyone to innovate; it requires a culture of trust. Equally fascinating, he revealed that people who regard themselves as innovators often have low rankings by others for “innovation”. In other words, the innovators are often less influencing than they imagine. He praised Agile processes as a way to encourage innovation.

Mark Loftus provided, incidentally, the most worrying comment of the day, when he stated, as a professional researcher rather than an academic, it would have cost him and his team a fortune to access the articles they wanted to read for their work. When potential users cannot access the content because of its cost, the academic studies might as well not exist – or users are tempted to bypass the payment system completely. Neither option benefits the authors.

Pooja Agarwal described very clearly initiatives underway at Bloomsbury Academic to improve diversity, both internal (staff diversity) and external (changing the balance of authors) with the obvious starting point that you have to measure before you can detect improvement. There is (or was until recently) no record of ethnic origin of Bloomsbury authors, for example.

Ian Mulvany of BMJ also talked about diversification, but this time of income, not ethnic origin. It was only in 2003 that BMJ Publishing became an independent entity, and even today, one journal, the British Medical Journal, comprises 15% of BMJ income. He described some good examples of diversification, for example how the job adverts at the back of the print BMJ gained a new lease of life in online form when the site was used to advertise NHS vacancies – in other words, if the traffic has gone elsewhere, work out a way to bring the traffic back, even if it means you are advertising someone else’s vacancies. The bigger opportunity at BMJ is the ability to track the healthcare professional using the BMJ resources, a goal shared by many societies.

A similar message was clear from Prathik Roy of Springer Nature. They are using third-party software companies to achieve more sophisticated text and data mining (TDM) tools, for questions such as “which genes are associated with this medical condition”. He also mentioned the increasingly popular way of doing commercial deals with seed-funded companies: the partner pays little initially to use the Springer data, and then if the partnership becomes successful, they pay more. This encourages innovation and recognizes that external partners can be essential for growth. Asked from the audience what he recommended for a small publisher wanting to make use of TDM, he suggested they partner with an external organization.

It felt to me that Kirsty Butcher of Manchester-based Place NorthWest came up with more initiatives in 20 minutes than most of the other presentations put together. This small property newspaper, which operates a controlled circulation (i.e. given away free to professionals in the sector), makes use of analytics to understand what customers are reading about, and then creates new content, podcasts, events, video tours of building sites, a quite remarkable range. When asked how they manage to innovate so quickly, Kirsty responded “we have journalists – they work very fast.” It was enough to make society and journal publishers highly envious.

Tasha Mellins-Cohen gave a rather defensive presentation about COUNTER, stating there is a major update to the software every five years (it seems that users have been complaining the changes were too frequent). My concern about COUNTER is not the innovation, but software project lifecycle. “We have a fabulous technical advisory group, who do not always appreciate the complexity of what they are recommending” was her description of the process, which did not inspire confidence. “COUNTER is just one staff person –   me.” That seems to me too small an organisation for the work required. My fear is that COUNTER is squeezed between several initiatives for measuring usage, and risks becoming less relevant as it tries to keep up with innovations such as Open Access and changes in funding, which continue to change the goalposts for metrics.

Steve Lodge of Emerald Publishing described their Impact initiative, a commendable effort to promote impact. For Emerald,  he stated, “Impact is our major headache”, particularly the lack of impact literacy in the sector – the understanding of how to identify what effect a publication has. This new subscription service for institutions sounds like a much-needed initiative, although he mentioned there are no immediate plans to provide specific measurements and hence comparisons. My guess is that measurements are what the institutions will be looking for, and they would welcome the chance to state their project, or institute, or department, has more impact than another one down the road.  

All in all, an excellent day, and plenty to discuss. Many thanks to NEC for continuing to present one of the most successful day conferences on the academic circuit.