A person on a laptop using artificial intelligence tool ChatGPT. Photo / Alex Cairns
An artificial intelligence tool that can generate essays, stories and poems and could potentially be used by students to cheat is proving a challenge — and an opportunity — for schools and tertiary institutes.
Educators spoke to the Rotorua Daily Post about plagiarism concerns and how the “exciting disruptor” ChatGPT can help teachers and students.
ChatGPT, released in November, is a chatbot created by US company OpenAI.
The NZ Herald previously reported the artificial intelligence (AI) programme had prompted some New Zealand universities to consider relying less on essays to assess learning and move towards in-person assessments such as oral examinations or presentations.
Toi Ohomai-Te Pūkenga head of kaiako [teacher] success and innovation Josh Burrell said it was taking a “positive” approach to the “really powerful” tool.
Burrell said it would benefit students and staff, but would also present “challenges” around the academic integrity of work.
Tertiary providers would have to “continually adapt” as technology improved, he said.
“This is a transformation in the way that society will operate moving forward. There’s not going to be many aspects of humanity that won’t be touched by AI,” he said.
“As educationists, we have a responsibility to ensure we are equipping our students to adapt to these technologies so they can make the best use of them in a positive way as they move into our workforce.”
Burrell said a team of staff Toi Ohomai had been working with faculty tutors to see how the programme could be used as an “efficiency tool”.
For example, ChatGPT could take a text-heavy PDF or PowerPoint presentation and develop a set of multi-choice questions in “minutes” - a task that would typically take a teacher “multiple hours”.
This month, Turnitin, an internet-based plagiarism detection service, released software that identified the use of AI-generated writing in academic work.
Toi Ohomai staff had also been “extensively” testing this detection tool, which he said was doing an “impressive job” of identifying any AI-generated content.
Burrell said Toi Ohomai had a comprehensive set of policies around academic integrity which included the expectation of submitting original work.
Students were expected to reference AI as a source, which would only “become more important” as its use increased, he said.
ChatGPT could assist students in research, producing revision questions and flashcards and translating content into different languages.
Most importantly, students needed to learn how to “critique the tool” for accuracy and use it “respectfully” in assessments, he said.
“We can’t allow our students to fall into the trap of just accepting the answers or the responses that AI gives us.”
University of Waikato pro-vice-chancellor of teaching and learning, Tracy Bowell, said the development of AI was a “fluid, fast-moving situation” and described ChatGPT as an “exciting disruptor” for universities.
Bowell said integrating AI into teaching required time, reflection and care. A team of staff had been putting the technology to the test since November, she said.
They were working on guidelines around how lecturers could incorporate it in courses, along with advice on designing assessments to make it harder for students to cheat.
Bowell believed it would lead to a rethink around how assessments were structured, particularly traditional writing exercises.
In paper outlines this year, students were told if they used AI to produce work passed off as their own, it would be considered cheating, she said.
Bowell said the university had a responsibility to teach students how to use AI “ethically” and “critically” and take these skills into the workplace.
“We can’t simply turn our back on the tool because we wouldn’t be meeting our responsibilities to prepare students well for the workplace.”
Digital equity, risks around academic integrity and reinforcement of bias were among her biggest concerns around the use of ChatGPT in academia, she said.
“We already have a wide digital equity gap,” she said.
“I wouldn’t want to replicate that with respect to access to these tools.”
Rotorua Lakes High School principal Jon Ward said he was concerned about how ChatGPT could impact “authenticity” in student assessment.
Ward said it was a matter being discussed among staff at the school. However, they had not yet developed any set rules around it, he said.
Clear guidance around the use of ChatGPT and other AI technology would “absolutely” need to be put in place in the near future, he said.
Rotorua Girls’ High School principal Sarah Davis said at the start of the school year, students were spoken to about assessments, meeting deadlines and plagiarism.
Using AI programmes such as ChatGPT was covered in this “general conversation,” she said.
Davis said teachers were familiar with their students’ writing styles so would likely recognise if it was not their own writing.
Assistant principal Sarah Riley said some staff were still “in the dark” about exactly how ChatGPT worked, but there had been very few instances of students using it for their school work.
“Usually … the writing is above what we have ever witnessed from the students.”
If teachers had suspicions about a piece of work, they could copy and paste it into the programme and ask it whether it had been produced by AI, she said.
John Paul College principal Justin Harper told the Rotorua Daily Post he was still considering how ChatGPT would impact students’ learning.
He was not for or against it, but said it had the potential to be a “disruptive technology” in the education sector.
“I used to be a mathematics teacher. When calculators started to come out, people were worried, but they have been integrated into the education system. I think that this is more disruptive than calculators.”
Ministry of Education strategy and integration general manager Pauline Cleaver said there were “opportunities and risks” in using AI in the classroom.
“Because AI can acquire bias, students may not be aware that the answers given by AI apps may be incorrect or inaccurate.”
AI’s ability to write “poems, stories, musical lyrics, and essays” may also require the focus of learning and assessment to shift, she said.
This could involve “students to be more creators of technology and have the ability think critically about its use as opposed to only consuming technology”.
Cleaver said blocking or banning AI was “not necessarily” the best long-term approach and the ministry would need to “strike a balance” to maximise the benefits AI offers.
It was developing guidance for schools, she said.
In February, The New Zealand Qualifications Authority (NZQA) provided a range of advice to teachers and principals around the use of ChatGPT and other AI programmes.
Local Democracy Reporting obtained the decision through the court.