The hype around ChatGPT is palpable: schools are racing to have staff meetings on it, share resources, and react quickly. After decades of being steadily and collectively formatted by a technology-driven economy and the resultant positivist ideology, peppered with our schizoid relationship with artificial intelligence (the pendulum swings from the feeling that we have to get ahead with artificial intelligence at all costs, to the other extreme of fear and reactivity), ChatGPT leaps out at us as from the dystopian sea of Deep Blue the “big data” craze and the robot Sophia. Remember the excitement and panic around Google search engines and Wikipedia and how much it would change everything? Here we go again … One can expect references to ChatGPT in conference keynotes for the next few months, speakers having grown tired of videos of robots or fun statistics about computing processing power.
And with these waves of advances in machine learning come the habitual rather hackneyed warnings of how this will change the face of employment, education and even what it means to be human. Such crystal-ball gazing can be quite serious, but it can also become ridiculous, always couched in norms and generalisations. There is the idea that educators should be doing something about it, changing something, getting ready for the future, or accepting that the future has already happened, and this needs to be done quickly and so on.
I don’t mean to denigrate the serious challenges and opportunities that technology creates (and always has as it has evolved historically), I write about it at length and seriously in Educating for the Twenty-First Century: Seven Global Challenges, but I do think that a pinch of salt here and there is not a bad thing either.
Here are my four thoughts on ChatGPT at this early stage:
- Artificial intelligence should be where thinking starts, not where it ends. (that phrase comes from my colleague Yoni Osman). Those who still hang on to the idea of banning technology in schools are fighting a losing battle. I once heard a rather sad story of a child being told off by a teacher for correcting him in class having accessed Google on her phone to verify what was said. Is the goal to access knowledge or to hide it? Curriculum has to be relevant, and the world we are living in is assisted by technology. Sometimes I wonder if cell phones have not replaced parts of our neocortical architecture. It might not be what we want to see happening to us as humans, but trying to create a romantic garden where there is no technology and algorithms such as ChatGPT are closed is absurd. On the contrary, algorithms save time. This does not mean that we should stop reading or doing research, it just means that such arcane intellectual and academic pursuits need to be accompanied with technology in more colourful ways. Students can use ChatGPT for certain in-class supervised tasks, they can ask it to give feedback on the occasional essay, but they need to use their faculties of discernment to discuss what is being created.
- The death of the search engine. One can imagine that ChatGPT will replace Google searches. Unfortunately, most users of search engines skim quickly through the first items that appear, so having artificial intelligence spew the answer out to you or create a narrative to your question instead is merely an accentuation of the same linearity. Discerning scholars might sieve through search engine responses, using their background knowledge and critical thinking to select the best sources, and they will have to do the same with ChatGPT, scrutinising it.
- The sterility of the artificial voice. As powerful as ChatGPT might be, and we should not underestimate its deep learning capacity (in a few months it will be much more powerful), when you read the texts it produces, somewhere, in the distant recesses of the prose, is a slightly flat monotonous machine-like sterility, resonant of the computer Hal’s metallic, depressed voice in Kubrick’s masterpiece 2001: A Space Odyssey. ChatGPT might sound human, but in that case, it sounds like a fairly dull human. And there are errors of course because of the automaticity of its understanding of semantics. Ultimately, expression is about style and voice, which is what I encourage in my philosophy students’ writing: give personal anecdotes, let the reader hear your voice. Perhaps it is a good thing that artificial intelligence is competing with us, it pushes us to bring out more vivacity, originality and personality in what we say.
- Plagiarism is only as big a threat as we allow it to be. One obvious reaction to this development in technology is to fear that students will no longer bother to write essays at home, or even design projects, since they will simply ask ChatGPT to do it for them. This merely reinforces the problem of extended pieces of homework. As soon as we ask students to do substantive pieces of work at home, there are a host of ways of plagiarising, getting others to do the work, using pre-packaged prompts and so on. I always have my students write every essay in class. At home they read and revise. Teachers who are fond of giving important and lengthy homework to students might have to rethink that strategy. And let’s not forget the Pandora’s box of assessment reliability problems that homework releases anyway, including issues of unequal work conditions at home: one of the reasons why schools exist in the first place is to create fair and equal learning conditions for all students. For coursework assignments, students might plagiarise and plagiarism-detection software will not pick it up. However, the irony of ironies is that ChatGPT cannot lie, so all we need to do is ask it if it wrote the piece of writing in question, and it will tell us.
All in all, the purpose of education is to live a better life, and this is achieved through the development of knowledge, aesthetics and virtues. Using artificial intelligence to accelerate processes and find solutions should not be a problem, but it will be our fault if we let artificial intelligence carve out the end of our thinking and not the beginning. All the texts in the world written by ChatGPT will not make us any wiser, and we have to teach our students to integrate words, ideas and values for themselves, for it is with this truly natural, inner language that they will define and understand the world, and the world will be as interesting as the depth of concepts are to describe it, no more, no less. As the great Austrian-British philosopher Ludwig Wittgenstein once said, “the limits of my language mean the limits of my world”. May those limits be the ones we choose, not the ones chosen for us by an algorithm.