Schools chiefs warn of ‘dystopian’ rise of artificial intelligence

Australia’s school curriculum chief has warned of a ‘dystopian future’’ for children struggling to separate fact from fiction, as artificial intelligence transforms knowledge.

Experts warn that AI could ­encourage cheating and peddle misinformation to students. Picture: AFP

Australia’s school curriculum chief has warned of a “dystopian future’’ for children struggling to separate fact from fiction, as artificial intelligence (AI) transforms knowledge.

As the nation’s education ministers draft the first rules for AI in schools, Australian Curriculum, Assessment and Reporting Authority (ACARA) chief executive David de Carvalho has called for greater focus on “facts and truth’’ in teaching.

He said students needed the “knowledge and wisdom’’ to ­detect lies, error, bias and deep fakes generated by AI.

The role of teachers as ­“authoritative sources of information, knowledge and wisdom” needed urgent buttressing.

“Our children are inheriting a dystopian brave new world and need to be equipped for the knowledge, capabilities and attitudes needed to renew and re-humanise the world,‘’ he writes in ­Inquirer today.

“In addition to reading, writing, numeracy and digital literacy … ethical understanding, personal and social capability, intercultural understanding and critical and creative thinking are going to be more and more important.

“But this focus cannot come at the expense of factual knowledge and an emphasis on truth.”

Federal Education Minister Jason Clare said a draft framework for using AI in schools is being prepared for his next meeting with state and territory counterparts in July.

“AI tools like ChatGPT can help you learn, but they can also help you cheat,’’ Mr Clare said on Friday.

Education ministers were briefed in February by University of Technology Sydney industry professor Leslie Loble, who served for 20 years as deputy ­secretary of the NSW Education Department until 2020.

Professor Loble called for ­urgent controls over the use of AI, and for children to be taught to question and control it.

“The time is now to set ­standards,’’ Professor Loble said on Friday.

“Schools need advice about what are high-quality (AI) tools built on good research.

“If we get the wrong tools and end up using AI in the wrong way, we could end up exacerbating learning gaps.’’

Professor Loble said AI would be expensive for schools, given that ChatGPT was now charging $240 a year for full-service ­accounts that could cost a school with 1000 students the equivalent of two teachers’ salaries.

“Schools are struggling with resourcing and time as it is,’’ she said.

“We need to set standards for what’s quality and provide ­resources to schools and teachers to be able to use it in the right way.

“You cannot assume these (AI) tools are accurate.’’

Professor Loble’s concerns were echoed by experts on the federal government’s National Science and Technology Council, who warned that AI could ­encourage cheating and peddle misinformation to students.

Emeritus Professor Cheryl Praeger, one of Australia’s leading mathematicians, called for a total rethink of teaching.

“I think that the whole issue of education and assessment will have to be rethought completely,’’ she said on Friday.

“I mean, cheating must ­become easier.

“There will be huge changes in the way we assess students ­because we can’t expect them not to be using tools which might be available to them.’’

Professor Praeger said ­students would need strong skills in critical and logical thinking to determine flaws in AI-generated solutions.

“I just hope that they learn (these skills) instead of just thinking that the computer’s going to do everything for them,’’ Professor Praeger said.

“I’ve had a look at some things that (AI website) ChatGPT throws up, which on the surface seem kind of reasonable, and then you see that there’s a huge gap in logic or in the argument.

“Students … need to recognise if something’s correct or not, but they really need to be able to discern and critique the logic of something.’’

A fellow member of the technology council, Griffith University molecular geneticist Associate Professor Jeremy Brownlie, said AI could generate misinformation accidentally or on purpose.

“While these systems might be acquiring information and then bringing it back, and students might be just taking it at face value, that is an issue,” Associate Professor Brownlie said.

“Misinformation is something that might be accidental, so it might relate to a bias in the data set that these AI systems are ­mining.”

The Australian has revealedleft-wing political bias in Google’s AI chatbot, Bard, which praises Labor Prime Minister Anthony Albanese as a “man of the people’’ while labelling Liberal leader Peter Dutton as “controversial’’.