ChatGPT: Can students use AI tools in college to pass?

As exam season begins, students may seek to use new artificial intelligence (AI) tools to gain an edge in assessments.

Universities have worked hard to understand what AI applications such as ChatGPT can do and introduce guidance on how to use them, but now they are being asked to teach students how to use them.

Academics at the University of Bath have examined the challenges and opportunities.

“Our first question was, ‘Can students use this to answer assessment questions?'” James Fern says of ChatGPT. ChatGPT is an online tool that allows you to answer questions in human-like language, such as writing essays and emails.

“For example, multiple-choice questions are handled very well.

“We definitely didn’t expect it to work. We were closer to 100%.”

But it struggles with more complex problems that require students to think critically.

An example from the final year evaluation reads: [a technical term, according to James] Overweight? ”

And there are signs of evidence that the answers given by ChatGPT were not written by students.

“At first glance, it looks very good. It looks very clearly written. The language looks very professional,” says James.

However, some statements are more like GCSE student statements than college students.

I have a habit of repeating the exact wording of the question “just written in a slightly different way” in the beginning and conclusion.

And when citing sources, they simply make them up, as is standard in academic research.

“It looks perfect. The author’s name is correct, the journal’s name is correct, and the titles all make a lot of sense. They just don’t exist,” says James.

“If you don’t know how large language models work, it’s easy to assume these are real references.”

About 6 months after ChatGPT went public, many students were wondering when they could and could not use it.

“I might be tempted to use ChatGPT, but I’m too scared to use it right now because I might get caught,” said one student walking between classes on campus.

“It is not yet clear what constitutes cheating on ChatGPT,” said another. “It’s cheating if you copy the entire quota from ChatGPT, but it would be very helpful to guide you.”

New advice from the Quality Assurance Agency, which reviews UK university standards, encourages students to develop AI skills that they can integrate into the world of work.

In September, we explained to incoming and returning students when and how AI should be used, and encouraged them to adjust their courses as needed.

Marketing lecturer Kim Watts calls it “another tool in your toolbox.” And some students in her department have already started using her ChatGPT this semester in a class that asks them to create a marketing plan.

“Students who don’t know where to start are encouraged to visit ChatGPT…and start playing with prompts,” she says.

“It doesn’t give them answers, but it can give them ideas.”

Kim demonstrates by asking ChatGPT to create her own marketing plan.

From creating a brand identity to using social media, we have a series of numbered points for you.

But Kim looked up from the screen and said, “This doesn’t work.

“To submit something like this is not detailed enough. It does not demonstrate learning or critical thinking.”

Neurodiverse and non-native English-speaking students will benefit most from ChatGPT, Kim said.

However, students who choose to use it are asked to submit ChatGPT’s prompts and answers as an appendix to clarify “how far have we come” from the chatbot’s answers.

Like most universities, Bath’s policies regarding ChatGPT and other AI tools are still being worked out. It is scheduled to be introduced from September.

The team then meets throughout the year to keep up with rapidly changing technology.

In the meantime, many staff members are once again setting up face-to-face proctored summer exams.

Dr. Chris Bonfield, head of the team that assists with the design evaluation, says the “default assumption” is that students shouldn’t be using ChatGPT this year. Also, expectations should be clearly stated by the staff if they allow it.

The pace of technological evolution poses a challenge for the university, but Barth quickly moved away from conversations about technology bans.

“This tool is not going away,” says Chris.

“We need to be involved not only in ensuring students have the skills they need for the workplace of the future, but also in keeping their degrees current.”

Professor Verena Rieser, a computer scientist at Heriot-Watt University who has worked in the AI ​​field for 20 years, says his students are “using AI in very creative ways,” but chatbots are It’s still in the early stages of development and can be used to generate false information in [a] When it comes to education, it’s clearly a very worrying scale.

ChatGPT’s previous model was never released because it was deemed “too risky,” she says.

“Like any technology, these tools carry real risks,” said its developer, OpenAI, adding, “We ensure that safety is built into the system at all levels. It works to “do”.

“We expect to see different types of ChatGPT from different companies soon, as well as more secure models that actually mitigate possible dangers,” says Verena.

“At the moment, I really don’t know how to stop the model from outputting false, toxic or hateful information. This is a big problem.”