When I studied at Berkeley in California over 30 years ago, we discussed the new phenomenon of artificial intelligence (AI) and its significance. The philosopher John Searle talked about the Chinese room – could you decide whether it was a machine or a human who answered at the other end? Patricia Churchland, popular in Cognitive Science and for the book Neurophilosophy (1986), argued against our fellow philosophy professor Hubert Dreyfus, who said that AI machines could not achieve human expert knowledge. The latter probably preferred Churchland & co. as eliminative materialists – where they considered our consciousness to be a computer.
Well, today, supercomputers and Big Tech have gained a capacity allowing AI to approach human capacity, replacing skills and communication in an increasing number of fields.
Among today’s ever-new chatbots with artificial intelligence – communicative apps we humans ‘converse’ with – much is now being written about the newer ChatGPT from OpenAI, created by the company Elon Musk helped start in 2015 (he developed self-driving Teslas, after all).
ChatGPT has gained over 100 million users within a few months – I am one of them. As an editor, I submitted a text from a writer and received responses to its weaknesses and possible improvements. And ChatGPT elegantly translated a text from English to Norwegian. And when asked, I got an assessment of the strengths and weaknesses of two competing video-cutting apps. Also: One of NY TID’s writers generated an article about Dag Solstad with ChatGPT in Norwegian.
Other visual programs such as DALL · E 2 from the same OpenAI or Midjourney (run inside the discord.com system) create images from sentences you write. For example, I received suggestions for several ‘photos’ of couples in a sailboat with wine glasses in the sunset and as defined – with dolphins in the background. Or see the author Maurice Blanchot (pictured), where MidJourney generated four pictures of him writing. Why? This will be revealed at the end of this text.
But what challenges do such new communication technology raise, based on huge databases and powerful collection algorithms (from the internet), and new ‘learning’ from the activity of millions of users? I can think of three, with their pros and cons:
Firstly: The thinking behind articles, books, speeches, research, art and homework will be diluted. Well, even if what has been expressed in text and thinking throughout the ages has always been somewhat parasitic on what others have written or said. But what happens when a teacher, editor or publisher can no longer tell much of a difference between what is delivered by a person or a machine? We’re not talking about ‘plagiarism’ here, are we? Some are now discussing being able to ‘watermark’ what ChatGPT produces to reveal who is behind it – but that is too complicated.
Secondly: What if ChatGPT can give you advice in Norwegian within a few seconds on requests that you would typically ask a lawyer, accountant, politician or psychologist? Some of these will be unemployed in the future – including authors and publishers with useful specialist books. And who needs an extended education as a lawyer (to read the law book?), an engineer (when machines calculate) or a cultural worker (when reports and expressions can be generated) – which lies behind such information or communication professions? (One employer said they got more efficient employees via ChatGPT, but it probably won’t be long before he starts reducing his staff.)
Third: Who will own or have copyright on information, texts, algorithms and images in the future – when ChatGPT and other such AI-based programs create new texts and images with fragments of other people’s work, databases and an internet full of letters, program codes and graphics? Will it be possible for artists and writers to get paid for individual work, and what about the income of large image agencies and publishers?
But ask yourself: Aren’t digital property rights both exaggerated as something ‘individual’ and neglected in terms of how technology, algorithms and practices are shared and copied today? Finished digital ‘products’ have no material scarcity and can therefore be easily copied and distributed – whether a report, an article, a picture, a work of art or a film. The owners can still track ‘illegal’ copying or use it online for a while. But today, lawsuits are only filed against established organizations or companies, as pursuing the ‘copies’ taken by private individuals or posted on small websites is not worthwhile. But with artificial intelligence, when the copy is a collection of fragments from several texts, images or entirely new expressions temporarily created by algorithms, this becomes difficult to pursue.
But ask yourself: Aren’t digital property rights both exaggerated as something ‘individual’ and neglected in terms of how technology, algorithms and practices are shared and copied today?
Are we really facing a major change in ‘the personal’ – in a way, the philosopher Roberto Esposito has criticized the term itself or our perception of what a person is in the book Terms of the Political (2008) and Third Person (2013). He criticizes the modern understanding of the individual as based on a false dualism between the personal and the impersonal. In modern society, individualism is put before the community, leading to political and societal crises. With the term impersonal community, Esposito proposes a community that recognizes our dependence, necessary relationships, and the shared life we humans basically have. He emphasizes here the ‘third person’ – as distinct from ‘I’ and ‘you’ – as something more humane and universal, as Simone Weil legally uses it. Or ‘the outside’, as Gilles Deleuze seeks immanence itself, the jointly created reality and language here among us.
Esposito bases this on a criticism that ‘persons’ have traditionally taken privileges, such as in the Roman Empire, where the head of the family ruled over both children, wives and slaves more or less as ‘things’ rather than ‘persons’ – who were at his disposal. This idea that someone stands above others, who can easily be killed, as they are ‘non-persons’, is also something the philosopher Giorgio Agamben has written about in his Homo Sacer series.
My point here is whether the copyright of a work should really be interpreted more as a common property – and that we should thus welcome ChatGPT, even if today it ‘steals’ an enormous amount of knowledge, artwork and linguistic erudition.
French Maurice Blanchot is also useful for this paradigm shift away from the person or individual focus: As the post-structuralist he was, Blanchot pointed out that the author himself should ideally disappear from the text – rather than the author using himself as the narrator, the text should almost play off itself with current events. The text should be «autonomous», and the author should use a «neutral voice.» The work could ‘speak’ for itself without an I-narrator or an I-you dialogue. Rather via a ‘third party’. The artist is considered by Blanchot instead as a «servant of the work» and not as its creator. This impersonality had consequences in literary theory – where instead, the reader’s role as interpreter was emphasized more.
And isn’t it precisely this devaluation of the ‘creator’ of a work that happens with ChatGPT when the answers produced by its artificial intelligence are mostly fragments from others that you stand on the shoulders of? Should this change the entire economic paradigm, to an open shared community, with open access to most things that can be copied? Isn’t it time that more people who deal with text, art and other produced works used by systems like ChatGPT are driven forward by social wages or citizen wages rather than struggling individually to recover money from copyright and licenses?
Only then, helped by ChatGPT and artificial intelligence, will we get what many have promoted – a real free creative commons, a creative community.