Artificial Intelligence and its Impact on Academic Publishing
Modern everyday life is beginning to look increasingly like scenes from science fiction movies, which previously seemed unrealistic: driverless trains are no longer novel; driverless cars are being tested and it seems they will soon be available on the market; and mobile touch-screen devices are already widely used. Now, newly standardized technology can respond to voice commands and even give responses.
Perhaps the oldest example of this in “civilian society” is Apple’s Siri, followed by the “OK Google” function on Android phones. Just a few years later, “Alexa” has entered the home: Amazon’s joke-telling, voice-activated device which sits in the corner of a room unnoticed. However, when “she” hears her name, she will jump into action, answering questions which are posed to test her general knowledge, switching on the light you have programmed her to control, and looking up train times. She can even be set to switch on and off the heating while you are out, or provide a welcome home message when you enter the house.
Other, perhaps less entertaining and purely functional automated devices now found in homes include robotic vacuum cleaners, which move about the room, keeping clear of any staircases until they return to “base” after a given amount of time, or before the battery runs out.
So, while homes are adapting to technological changes, the next question is how artificial intelligence (AI) will be used in the workplace.
AI in the working environment
Throughout history, technological advances have triggered shifts in industry and agriculture. Industrial and agricultural revolutions of the past are perhaps more easily defined than the shifting climate in which one lives; however, we are constantly faced with the threat of robots taking human jobs, and in today’s cultural and economic climate, new technology catches on very quickly. School curricula and higher education courses are quickly adapting so that new generations of computer engineers and technicians can be trained who will fit into and even help shape this brave new world.
The translation industry has already seen patterns of decreasing quantities of work needing traditional translation, and increasing quantities of texts that simply need post-editing: they need a human touch to ensure that computer-generated translations are fit for purpose.
The impact of AI on science and academia
Just as Alexa can scan the Internet in search of answers to our questions, there is already software available which scans large databases to carry out plagiarism checks on new manuscripts in the editorial process. At the moment, manual input is needed (as well as a paid subscription to the more comprehensive services), but with the advancements in voice-activated technology, it may be only a matter of time before this can be done with minimal human input.
This is precisely the debate discussed in the following article on Scholarly Kitchen: https://scholarlykitchen.sspnet.org/2017/03/22/living-with-an-ai-a-glimpse-into-the-future/. The author, David Smith, suggests that this would greatly enhance the peer review process: automated “pre-emptive” plagiarism checks could be carried out.
Further to this, is it possible that the entire peer review process could be automated, with no longer any need for human interaction?
An entirely automated peer review process?
The short answer, fortunately, is no.
Google Translate is proof that machines can learn grammatical and linguistic rules (rule-based machine translation), and statistical probabilities in language (statistics-based machine translation). In other words, it is easy to teach a programme, such as Microsoft Word, to know that “we is” is incorrect; it should be “we are”. Equally, machines can be taught that “open access” is a more likely collocation than “accessible access”, or “open admittance”.
That is not to say, however, that alternative phrases are never used, or that we always want to avoid bad grammar – after all, it provided a convenient example above. Computers are, therefore, not correct in every situation.
Why is this fortunate?
Because it means that we humans are not entirely replaceable!
Peer reviewers themselves could never be replaced by machines: humans are needed to review language competence, and because machines are only as good as the people who programme them, they could never keep ahead of scientific research. Humans are therefore needed to evaluate and provide feedback on manuscripts, as well as to feed information to computers to help them improve. Likewise, administrators handling research manuscripts will continue to be necessary for dealing with the unexpected: answering questions and managing projects, etc.
So, while computers may play an increasingly useful role in editorial and peer review processes, there are some things that must, for the foreseeable future at least, remain “old fashioned”. Arguably, the peer review process could be improved, and perhaps AI will speed it up somewhat and help it to evolve. However, until computers develop a true “intelligence”, only humans can react to situations which are out of the ordinary, and confirm whether research is innovative and of a high standard.
Additional help and support
Photograph: Alamy Stock Photo