AI Tools in Assessment: Should they be used?

by | Aug 2, 2023 | AI

Krishna is currently a postgraduate student of MA Digital Technologies, Communication, and Education at The University of Manchester. Her interests have been around education and technologies, especially how to enhance learning outcomes by incorporating technologies. 

As artificial intelligence (AI) emerged recently, she am willing to know how far AI can impact education. She was involved as a facilitator of the discussions had by JISC on AI in Assessment for when UoM hosted these sessions in April, and this blogpost is one of the insightful outcomes generated from it. I hope this piece of brief writing can offer a big picture and summary of how the discussion went. 

 

 

The field of artificial intelligence (AI) has been with us for years but rarely experienced in daily life until the end of 2022 when ChatGPT was publicly deployed. The emergence of ChatGPT has caused a disruption in many sectors, including education. Since its release, ChatGPT has reached 100 million users by January 2023. This means that many people have started incorporating this tool into their lives, which is probably also used by students in completing their assignments to some extent.

As a digital enabler for further education in England and higher education in the UK, Jisc UK is preparing themselves to deliver responsible services regarding AI for education, especially in tertiary education. In response to the situation, Jisc is gathering university students around the UK to collect and listen to students’ voice about their perspective once AI is embedded in assessments or education in general. Their voices are summarised in this post.

To date, students have been utilising different AI tools in different aspects of their lives, including their academic life. From the discussion, it is that ChatGPT and Quillbot become the most widely used AI tools to help them with assignments. While ChatGPT proves to be immensely helpful in explaining difficult topics by simplifying them into digestible bits and can also provide programming code, among other things. It is important to note that ChatGPT may generate incorrect information. For example, code responses from ChatGPT are not always accurate and functional. Similarly, when asked to give references, AI tools might fabricate them, despite appearing convincing. Hence, critically evaluating its responses is essential.

In addition, Quillbot’s contribution to students’ assignments is comparable to ChatGPT’s, as this tool helps students in paraphrasing a text. Interestingly, the Quillbot tool is not only employed by international students – those who speak English as a second language – but also by native students.

Regarding AI tools and its uses, students are aware of their implications on originality and academic integrity. It is evident that copying and pasting from AI-generated answers might taint the integrity of academic work. However, it is worrying when AI paraphrasing and grammar-checking tools are considered AI plagiarism. For many students, these tools are helpful in organising and structuring their thoughts into formal and scholarly language. Because those tools are not the ones who ‘do the work’ (make the paragraph), instead, it is still the students who ‘do the work’ where AI is an assistant in this case. Thus, students are expressing that need for clear guidelines regarding the use of AI in assessments; whether it involves using AI to assess the assignments or using it in our assignments, both should be under clear and transparent regulation. This could mean adjusting the marking processes (Furthermore, the discussion also led to an issue where AI literacy might be an important skill to be embedded into students to provide an overview of the ethics, as well as the boundaries in using AI tools for academic purposes.

In summary, AI tools can be embedded into assessments, yet it needs to have clear guidelines. Since this development of AI tools is potent to be greatly evolving in the upcoming times, educational institutions will have to regulate it mindfully when considering to embed it in the assessments. Not only the drawback such as issue on academic integrity, but also on the advantage that students can get bridge the gap of English fluency, especially in terms of academic writing.

AI tools indeed allow students – and everyone – to do some tasks easier, but one thing to keep in mind is that AI tools are trained based on the data that we provide, which also has misinterpretation and biased views. If we are unable to trust people a hundred per cent and always questioning about others’ perspective, then there should be nothing more to expect from AI tools; it is a tool. Worry not, this piece of writing is an original product of human creativity and criticality, not the AI… or is it?

 

Related Links:

0 Comments

The University of Manchester Institute of Teaching and Learning logo

TEA is a place where we can share practice and experience around teaching and learning with colleagues, students and the wider public – how we Teach, Explore, Apply

We’d really like to hear your responses to the stories we feature – if you want to respond, please email the ITL team at teaching.learning@manchester.ac.uk, or the University’s Teaching and Learning Online Network (TALON) on Yammer is a great place for you to carry on the discussion of any practice you find really interesting.