New technological developments are sometimes received in very different ways. Some people are positive and see all kinds of opportunities, while others are concerned and mostly point out dangers. High expectations often fail to materialise, dreaded dangers are curbed in practice, and the technology finds its place in the end , quite often in unforeseen ways.
At the moment, there is a lot of attention for a development that has quickly become available to many thanks to a clever marketing strategy. Based on more or less specific search queries, you can generate ‘free’ texts. These texts are created by mixing a large number of existing texts in such a way that makes it seem as if new text is being created. Of course, the service is not really free: anyone who uses this system is providing knowledge and data to the company behind this technology. From an ecological point of view, text generation is also not free: each request requires a lot of energy.
By now, I have become accustomed to the fact that many people have no problem giving away their data to mostly American and Chinese companies that do not fall under European legislation. But I am still amazed at the ease with which many people once again trust a company that stores all data beyond the reach of Dutch legislation, thus helping to improve the system. In the foreseeable future, this system will probably become so good that the company will start charging money for it. In what way that will happen is not yet known – will universities soon have to pay to check whether a text was created by AI and by whom? Will we soon have to pay to make our texts reliable, or unique? I guess we’ll have to wait and see.
What role will text-generating AI play in teaching and research? Those who have texts generated to avoid having to think for themselves or to have sufficient knowledge at their fingertips may be in for a rude awakening. After all, if you do not master a subject and have a text generated about it, you will hardly be in a position to properly assess the result for quality and reliability.
Dealing critically with knowledge and generating new knowledge from there is at the heart of academic education. This takes effort. But without that effort, development stalls. Of course, you can get a good grade for an AI-generated essay or thesis, but the only person you are truly disadvantaging in doing so is yourself. If you get your degree through fraud (because that is what it is), you will be at risk for the rest of your life. After all, a company that is out for profit will know how you generated that text. Worse, anyone (including yourself!) who thinks you possess certain academic skills based on your degree will be rather disappointed once you have to put those skills to use. This is about your personal academic training. At a fundamental level, I assume that students study in order to grow and develop, and not just to get a degree.
Still, let me end on a more positive note. I think that most students and staff members will want to use this new technology safely, appropriately and fairly. Not to pretend that they created the AI-generated texts themselves, but to get support and improve their own skills by using this kind of system. If such a system should become available, I will gladly use it. It is bound to improve the readability of my columns!
Disclaimer: This text has been produced by human intelligence. Artificial intelligence was solely used to check facts.
Han van Krieken is rector magnificus of Radboud University