One of the great myths surrounding technological innovation is its promise to create new things. In this way, it seems as if ChatGPT was the last expression of humanity’s development towards the highest in history. After the Google search engine altered the way we access knowledge, creating a kind of universal digital library, the Microsoft chatbot prototype would have come to reconfigure our relationship with education, the world of work and radically alter human relationships. .
Apart from the fact that the logic of ChatGPT is an heir to the times of the Cold War (the comparison of patterns, designed to predict and react in war environments), as is the case with the search engine or the GPS system of Google or other technologies like the iPhone, the obsession with classifications of apparently rational ideas illustrates a paradoxical event: the irrationality behind the way we understand technique.
On the one hand, because it tries to solve the problems that the “previous” digital world had created, which ends up creating more problems. It is what Evgeny Morozov has called solutionism: since this artificial intelligence has been trained with billions of pages and information from the Internet, which until recently we agreed was an ocean of fake news, many of the answers it delivers are little wise. Although it has capabilities to generate text, this tool represents something of a way of institutionalizing disinformation, although now with plugins and sophisticated prompts.
But, does anyone think that this will be the technique that history books remember as the library of the 21st century in the same way that, for example, similar cultural institutions created in El Escorial by Felipe II were? The problem is that ChatGPT tries to solve questions about our society that we should be dealing with in a different way. For example, in relation to education, is the problem that a machine is capable of creating dissertations automatically or that the public education systems (which have suffered the cuts first, and the googling later) are still unable to enjoy their own innovative digital infrastructures?
The problem is not that students have more tools to copy, but rather that education does not have a teaching and learning model that gives priority to the acquisition of analytical and synthetic skills over the domain of information. As the philosopher Roberto Mangabeira Unger expressed, the question is how to create tools that help people reach a certain selective depth to the detriment of encyclopedic superficiality in the treatment of content, or that prioritize collaborative work (among students, teachers and centers ) instead of individualism and authoritarianism in the classroom. To approach each topic from contrasting, truthful points of view, we need artificial intelligences very different from ChatGPT: machines trained with other data sets, designed to learn or discover new authors and pieces of knowledge, not to automate an outdated teaching model.
Something similar happens with other fields. In the book “Shitty Jobs”, the American anthropologist David Graber displayed a theory where he affirms that the existence of jobs without any purpose has corrosive effects on society and becomes psychologically destructive. So why do we create artificial intelligences that try to solve this problem by automating such jobs, instead of creating machines to reduce the workload, distribute the derived capital gains through some form of universal basic income, as proposed by Francesca Briaand allow human creativity to express itself freely?
Evgeny Morozov argued that ChatGPT is not intelligent or artificial: it draws its strength from the work of humans, be they artists, musicians, programmers or writers whose creative and professional production is appropriated in the name of saving civilization. Given this reality, how do we reprogram artificial intelligence to make human intelligence visible, capable of creating art, cultural fictions or new stories, instead of spending huge amounts of money and energy resources on data centers and machine learning models? Wouldn’t it be smarter to rethink the digital library of the 21st century following sustainability criteria, creating tools to shape futures that we would like to live, with knowledge about experiences that have been successful in the past, instead of something like automating global warming while raise doubts to a bot?
Ekaitz cancels He is the author of ‘Digital Utopias: Imagining the End of Capitalism’ (Verso Libros, 2023).