The new report, titled AI Use In Schools: Taking Action, draws on a survey of more than 4000 NSW teachers and school leaders across government, Catholic and independent schools, and reveals widespread concern among educators about the routine use of AI by students and its impact on learning.
About half of secondary teachers reported that students use AI for schoolwork, and the same amount are worried that they do not know how to prevent instances of plagiarism or cheating.
Eighty per cent of educators whose students use AI for schoolwork are worried about AI’s future impact on education.
“Like most things in education, and in ed tech, there’s a lot of people pushing it, and a lot of people pushing against it, but the research around the use of AI by students is actually pretty clear – it’s not good,” the study's co-author Dr Ben Jensen says.
The CEO of education research and consulting group Learning First says the cognitive outsourcing that occurs is not learning.
“What’s striking is just how much when teachers talked about students' use of AI, they echo the high quality research almost to a ‘tee’.
“One of the quotes in the report is ‘students are not using it as a learning partner, they’re using it to do the thinking for them’ and once we’re at that stage, then we’re talking about real learning loss.”
Jensen says what has really struck him about the report’s findings has been the comprehensive nature by which teachers and school leaders are across the issue, how widely they can pinpoint exactly what the problems are and how much that is reflected in robust research.
“At a policy or a system level, the debate has too much been on ‘this is something we have to deal with in the future’, so what we’re really trying to do is say ‘no, this isn’t something in five years' time we’re going to help schools with, we’ve got to help schools right now’.
“There is little evidence of the benefit of AI to student learning but clear research on the harm it can do.”
While the business world and many employers are calling for students to be more ‘job ready’ and for AI to be more deeply entrenched in the curriculum, Jensen says this is a trap that the sector must not fall for.
“There’s no doubt that AI is going to be a part of people’s work in the future, but that’s not a reason for putting it into primary and secondary schools,” he says.

“There is little evidence of the benefit of AI to student learning but clear research on the harm it can do. We must act now,” Dr Ben Jensen says.
By the time most students get into the workforce, AI practices will have changed completely, he notes.
“It’s the same rationale for coding we have heard for the past decade – how we must teach our kids coding. AI has come along and put virtually every coder out of work.”
A child currently in primary school is going to enter the workplace or in an office environment in about 2045.
“I feel very confident in saying that none of us know what AI practices or AI tools they’re going to be using in 2045,” Jensen says.
AI is used in the workplace to complete certain tasks – but that’s not what we’re doing in schools, he argues.
“The point of schooling is about learning, and so it’s a really different thing we should talk about.”
That’s not to say AI is going to be such a part of our world that we should stick our heads in the sand and pretend it doesn’t exist, Jensen says.
“If you’re in further education, university, post-secondary education and you’re say six months off, then, yes, start to learn some of those AI tools.
“Generally we’ll have to educate kids ‘this is what AI is, this is how it can be used, we think it’s not good to use it this way’, so in that sense we definitely need to train kids – but I disagree with that notion of ‘we should train kids to use AI and just live with a negative impact on learning’ just because that’s what’s happening in workplaces now'.”
The report calls for coordinated leadership at every level of the education system to address both the dangers and potential benefits of AI.
It advocates a push back on unfounded claims for AI’s use in schools, and an immediate restriction on students' use of AI.
It also calls for a review of senior secondary school assessment and steps taken to prevent AI from reducing confidence in assessment and qualification systems.
It also calls for the development of assessment programs in primary and lower secondary years that address AI plagiarism and cheating.
The report found that about three quarters of teachers are using AI to support their work, mainly to develop curriculum resources.

Research on the use of AI tools like ChatGPT for creating materials like lesson plans suggests “we're not looking at high quality curriculum resources”, Jensen says.
“The research has shown that if you use a general purpose AI, such as ChatGPT, the curriculum resources, the lesson plans etc, that you end up developing most often are pretty low quality ... we’re not looking at high quality curriculum resources,” Jensen laments.
The risk to the integrity of senior secondary assessment requires the most urgent attention, he says, with many assessments in these year levels susceptible to AI-related plagiarism or cheating.
“In the upper secondary year levels, whenever we’re talking about HSE high stakes school-based assessments, or the equivalent, we must have very strict guardrails around this and I think that what that means is we’ll probably have to move towards supervised in-school activities for these assessments,” he shares.
In lower secondary, schools need to redo or reimagine their assessment approach and develop a program that can generate enough trustworthy evidence to judge and report on student progress, while preserving rich learning experiences.
We need to progress carefully, the learning expert emphasises.
“Those long essays and larger projects have some great beneficial educational value, so we don’t want to end up in a world where we say ‘from now on all assessment has to be at one extreme, you know, short supervised tests in classroom’.
“We’ll need new assessment plans or assessment schedules in schools, where we actually have a range of activities, identify which assessment tasks are and are not more susceptible to AI plagiarism, and try and get a balance between the two.”
Jensen is cautiously optimistic about the road ahead.
“I switch between cautiously optimistic and absolutely horrified,” he admits.
“If I want to be glass half-full, in my mind this offers up the opportunity to place a greater emphasis on what we know are really great, rich learning experiences for kids.
“We see that teachers are already taking steps to deal with AI and what they told us is, they’re doing things like, if someone does an essay, they’re questioning them about that essay and talking to them about their methodology.
“They’re complementing written work with oral examinations, where people have to talk about the project they’ve just done.”
An emerging trend appears to be teachers placing greater emphasis on paper and pen, rather than having students stuck on screens all day.
“We’ve been heading down a road with ed tech that I don’t think is good,” Jensen concludes.
“One of our problems is, we’re not doing enough reading and writing, and we’re not doing enough reading and writing of different styles of text.
“When I talk to teachers and school principals, they’re more generally worried about all the various things we’re getting kids to do online for this or that.
“If this pushback against AI continues, and we start to move away from education technology, then I think that could have real benefits.”