Jo Raphael: Artificial intelligence should be embraced, but with caution

AI might have its benefits but, as it said itself, caution is needed, writes Jo Raphael. Photo / 123rf

OPINION

As a Gen-Xer I apparently have the ability to straddle two worlds: growing up without internet technology and being around at the emergence of its popularity and evolving and adapting as fast as the technology has.

That’s why I’ve become fascinated with artificial intelligence technology, target="_blank">ChatGPT.

Playing around on it the other day, I asked it to give me three sentences on the uses of Chatgpt — and here’s what it told me:

“ChatGPT can be used for a wide range of natural language processing tasks, such as text classification, sentiment analysis, and language generation.

“It can also be used to build chatbots, virtual assistants, and automated customer service agents that can interact with users in natural language.

“Additionally, ChatGPT can be used for educational and research purposes, such as language modelling and understanding, dialogue systems, and machine learning.”

That’s all well and fine, ChatGPT, but what are the risks?

This is what it had to say:

“As with any powerful technology, there are risks associated with using ChatGPT.

“One of the main risks is the potential for bias, as the language model is trained on vast amounts of data, some of which may contain biased or harmful content. This could lead to the language model producing responses that are offensive or discriminatory.

“Additionally, there is a risk of misinformation being spread through the use of ChatGPT, as the language model may generate false or misleading information if it is not properly trained or monitored.

“Finally, there is a risk of privacy breaches if sensitive or confidential information is shared with ChatGPT.”

Okay — ChatGPT does warn you of the potential downfalls when you sign up for its use.

But there is the additional worry that students could ask it to write their assignments — and we put this to Bay of Plenty educators this week and their responses were surprising, but in hindsight logical.

It means educators will need to adapt and embrace this technology, with one calling it an “exciting disrupter”.

Toi Ohomai-Te Pūkenga head of kaiako (teacher) success and innovation Josh Burrell says it is taking a positive approach to the “really powerful” tool.

Burrell says Toi Ohomai has academic integrity policies that include the expectation of submitting original work.

Students are expected to reference AI as a source, he says.

Some have mentioned when a student turns in a piece of work they suspect is written by AI, they can tell — “Usually the writing is above what we have ever witnessed from the students,” Rotorua Girls’ High assistant principal Sarah Riley says.

Also, there is an online tool they can use to find out if it has been generated online.

I, however, like the idea that teachers move away from asking students to write swathes of text that may or may not be artificially generated.

I believe the use of more oral examinations or asking students to make presentations and explain their work, to assess students would be beneficial.

They are asked to use their own words, their own ideas, from their own brain.

AI might have its benefits but, as it says itself, caution is needed.

If someone tries to pass off AI work as their own, it’s good to know that educators are realising this and preparing for it.