Trust, but Verify: Students’ Reflections on Using Artificial Intelligence in Written Assignments

This article focuses on students’ reflective assignments (n = 26) on using AI chatbots in the short literature review writing process. The opacity of text-generating AI software is impenetrable, yet the flaws in the output provided by AI chatbots sooner or later drive users from trusting to verify the information. The affective component of using AI chatbots is noteworthy as students experiencing time stress or lack of ideas need a ‘helping hand.’ However, with too few experiences in using AI chatbots, they do not anticipate that seemingly fluent text contains hallucinated references and false or machine-like content. In that case, the need to verify the information comes by a surprise, as the time initially reserved for the home assignment is no longer sufficient. Practice and peer-to-peer discussions about AI chatbots in the framework of information literacy related courses can help reduce problems related to using AI chatbots.