IN BRIEFNearly 80 per cent of Australian children are using AI assistants every day.Experts warn an overreliance on AI could lead to cognitive decline and skills loss.

Artificial Intelligence is becoming increasingly common in classrooms and workplaces across Australia. But, could using the technology make us dumber?

The AI assistant, Gemini, said using AI could improve a person’s memory and recall, problem-solving skills, and efficiency. It claimed that “cognitive scientists view AI as a bicycle for the mind”.

“[It] handles routine cognitive labour, freeing up mental energy for higher-level creativity and strategy, provided users remain actively engaged,” Gemini said.

However, former chief scientist of Australia and neuroscientist, Dr Alan Finkel, says otherwise.

“I don’t think your brain, or my brain, needs a bicycle to orgment it,” he told SBS News, adding that the AI assistant provided a “self-serving comment”.

Your trusted source for staying up-to-date with the world around you. Get free daily news updates and analysis, straight to your inbox.

“[It’s trying] to convince us that it is there to help us achieve. Whereas, in this respect, it is there to replace what we do,” he said.

Neuroplasticity expert and director at the Wicking Dementia Research and Education Centre, Emeritus Professor James Vickers, likened AI assistants to a “self-driving car”.

“You no longer need to think about how you get from Point A to Point B, you just put yourself in the car and it gets you there,” he told SBS News.

Vickers said there could be cognitive consequences from the use of AI to solve problems.

“The brain is very much a use it or lose it type of organ,” he said.

“If you don’t keep using it to tackle problems, then what you’ll see is a withering of those functional connections that allow you to respond. The more you offload that capacity to AI or a machine, the less capacity you have to do that.”

Almost 80 per cent of Australian children are using AI

In August, the eSafety Commissioner released a report that revealed that 78 per cent of Australian children use an AI assistant.

ChatGPT was the most popular, with 70 per cent of children using the assistant. Gemini and Copilot were each used by 20 per cent of the cohort.

The study found that 20 per cent of children used an AI assistant daily and 28 per cent used it a few times a week.

Over 90 per cent reported using the technology to find information or support their learning, 70 per cent used it for writing or editing and 68 per cent for entertainment.

READ MORE

Half of the surveyed cohort reported using an AI assistant for “personal or social reasons”.

There’s little research on what daily use of AI assistants does to a child’s neurocognitive development. However, Finkel says there is global evidence that children who use it for their homework suffer lower learning capacity and grades.

Collaborative research from the University of Hong Kong and Stockholm University studied 13,500 school students who used AI to assist with their homework and 13,500 school students who didn’t.

The two-year study found that while the students who used AI got better results for their homework, they scored 20 per cent lower in in-person classwork, than children who didn’t.

It’s a reality that’s already showing up in Australia’s national standardised school test, NAPLAN.

NAPLAN tests the reading, writing and numeracy skills of 1.2 million students across Years 3, 5, 7 and 9.

The latest results from August 2026, found that reading and writing skills were decreasing across every age group — with one in three children failing to meet NAPLAN benchmarks.

Finkel said that students who are turning to AI for help are failing “to really use their brain”.

There was increasing tension between on-paper results and learning, he added.

“Somebody who’s written a school essay as a draft or in dot points could say, “Listen, Claude or ChatGPT, can you take my dot points and turn it into a 2,000-word essay?”. That person has then failed to really use their brain to consolidate their thoughts into an elegant expression,” he said.

General practitioner and associate professor at the University of Melbourne’s Centre for AI and Digital Ethics, Dr Grant Blashki, said students should use generative AI as a “sparring partner, not a stunt double” and recommended educators focus on processes over outcome.

“The dilemma for us as educators … is if we are just giving students assignments that they can throw into AI, what are they learning from that? Reading an essay that they’ve just put into ChatGPT or Claude, what have we learnt there?”

He said he’s adjusted his teaching methods to combat this.

“We’ve changed our assessments now to what we call “assured learning”. As well as an essay, students have to do a one-on-one interview with us face-to-face, or on Zoom.”

What about educators or professionals?

Blashki uses AI regularly as a doctor. With patient consent, he uses an AI tool to transcribe medical appointments.

While it saves him time and lets him focus on more complex tasks, there’s a catch.

“When you’ve been using AI … it can make mistakes. It can so-called hallucinate or make errors, or hear something wrong,” he said.

“But when it’s mostly getting things right, you can get a bit dependent on it.”

This can lead to “skill loss” in seasoned doctors.

“For new doctors, if they’re already going into the medical system where AI is embedded, or they get reliant on it, they don’t learn to read the nonverbal cues and the gut feeling you get from practising medicine,” he said.

READ MORE

Blashki, who co-authored the book Artificial Intelligence for Better or Worse, said the AI debate is very complex.

“I don’t think it’s as simple as AI bad, using your brain good,” he said.

“We can be intelligent about the way we use it.

“As our successful young Australians go out into the workforce and the world, if we can make them smart with how they use AI, to my mind, that’s a great investment.”

For the latest from SBS News, download our app and subscribe to our newsletter.