As the time for exams begins, the question of whether artificial intelligence (AI) is able to write, for example, a bachelor’s thesis instead of a student, becomes particularly relevant, writes the BBC.
Universities are trying to understand what AI apps like ChatGPT can do and how to develop guidelines for working with them. University of Bath researcher James Fern says about ChatGPT: “Our first question was, ‘Could this be used by students to answer our assessment questions?’ The researcher notes that the app handles multiple-choice questions well, even surprisingly well.
However,
when it comes to more complex questions that require critical thinking, the chatbot is no longer so infallible.
Answers to such questions show several signs that the answer is from an AI rather than a live student. Fern states: “At first glance, it looks very good – it looks very clearly written, it looks quite professional in its language.” Nevertheless, there are features that make the answer more akin to the judgment of a high school pupil than that of a field student. ChatGPT answers use a slightly modified repetition of the question phrase in the introduction and conclusion. When you need to refer to a source of information, as is common in academic works, the chatbot simply creates them.
“They look perfect – they’ve got the right names of authors, they’ve got the right names of journals, the titles all sound very sensible – they just don’t exist.”
Since ChatGPT became available to the general public about six months ago, many students have been confused about when they can and can’t use it.
British Education Secretary Gillian Keegan said on Monday, the 8th of May, that AI is already making a difference in schools and universities, and indicated that it could help teachers plan lessons.
Both Bath and other British
universities are still working on the development of rules for the use of ChatGPT and other AI tools in the study process.
It is expected that the new rules could enter into force in September, and after that regular meetings of the responsible persons are planned to adapt the rules to the situation. Meanwhile, the exam session is approaching and the university’s academic staff is organizing the usual face-to-face tests.
Geoffrey Hinton, who is considered the godfather of artificial intelligence, quit his job at Google in early May and warned of the growing dangers posed by AI and chatbots. Verena Rieser, a professor at Heriot-Watt University who has been working with artificial intelligence for two decades, said that while her students are using chatbots in creative ways, they are still in their infancy and, when it comes to education, can be used to spread misinformation on a scale that is very disturbing.
OpenAI, the developer of ChatGPT, says that like any other technology, these tools also carry real risks, but the company is working to make its systems secure.
Read also: Godfather of AI warns of dangers