Experts cautiously optimistic about expanding AI in education, health sectors

Experts are cautiously optimistic about the government’s interest in developing artificial intelligence, but they would like to see a considered and thoroughly researched approach.

Digital Minister Judith Collins wants to start with the health and education sectors. She says the new technology could be used to assess mammograms and teach children.

Dr Michael Johnston, a senior researcher at the New Zealand Initiative, said many in the education sector hoped the technology would level the playing field, as research showed students who received one-on-one tutoring performed better (two standard deviations higher) than those who did not.

“Two standard deviations take you from the 50th percentile to the 95th percentile, so that’s huge. The problem, of course, is that we can’t afford to have a one-on-one tutor for every kid in a school, so if AI can become some kind of substitute, that could have a significant impact on educational equity.”

However, Johnston warns that the technology should not be used as a substitute for learning basic skills.

“The use of any technology, and in particular artificial intelligence, in education must not make it difficult for children and young people to acquire the knowledge and skills they will later need for further learning,” he said.

“A good example is early math, learning arithmetic. If we let kids use calculators to add and subtract before they can do those things themselves, before they have mastered those skills, then anything that comes later that has to build on those skills is going to have a shaky foundation.”

He added that the rise in the use of artificial intelligence also raises questions about whether young people will spend even more time in front of screens.

“You’re not only putting them at risk of not developing social skills, but in fact many children now start school with insufficient oral language skills, not only to learn to read and write, but also to access what’s going on in the classroom, to understand what’s going on. That’s a real problem.”

Artificial Intelligence Can Relieve Teachers of Administrative Duties – Academic

A leading artificial intelligence researcher says using AI in education could deliver significant benefits for teachers and students.

Mahsa McCauley, Senior Lecturer at Auckland University of Technology and Director of the AI ​​Forum, said: Control point This technology could relieve teachers of administrative duties.

She added that AI could take care of repetitive, boring tasks so teachers could focus on things like spending more one-on-one time with students.

She added that there are still many unknowns about the use of artificial intelligence in education.

She added that a potential benefit of using AI could be creating personalized learning experiences for students.

“So each student can have their own customized curriculum that is really based on their needs and their pace, and the other thing is access to an intelligent learning system that is available to the student, an assessment system that is available to support the student 24/7.”

Thanks to this, each teacher will be able to have their own virtual assistant, she added.

However, she added that AI also brings with it certain risks that must be addressed while remaining open to the “transformational potential of this technology.”

McCauley said parents, teachers and children need to be trained in how to use AI before the technology is implemented.

“We must advocate for responsible and equitable use of AI in the classroom.”

Artificial intelligence could be a ‘major breakthrough’ in healthcare – general surgeon

In health, Collins wanted AI to take on tasks like processing mammograms. General surgeon and University of Auckland associate professor Dr Matthew Clark said it would be the “biggest game changer” for the health sector in generations.

“One of the biggest challenges at the moment is the different languages ​​that people speak in Aotearoa. Perhaps using chatbots in a person’s native language to improve health communication, to convey information, will be an early win.”

However, he agreed that the technology would still require human oversight.

“You still have to approach this with caution because we know we have to make sure the technology is safe and people are not being misdiagnosed or misdiagnosed at these very high risk levels.”

Clark said that currently, AI is being used in healthcare on a small, isolated scale, so it is difficult to know how it is already being applied.

“One of the things I hope for in the future is a clearinghouse of technology and AI so that everyone knows what’s going on and those involved can take advantage of the winnings we generate and quickly distribute them across the motu.”

Allyn Robins says AI guidelines needed at government level

Allyn Robins, head of AI think tank Brainbox, said there was reason to be enthusiastic but warned that people should not get carried away by the “peak of the AI ​​hype cycle.”

For example, Robins said the technology is not yet suitable for purposes such as tutoring.

“These AI systems, certainly the ones that (Collins) seems to be talking about, do not think in the way that the term ‘artificial intelligence’ often suggests,” he said.

“They all do essentially the same thing, which is to put words together in the way they’ve been put together before, and that means they’re not fundamentally designed to share facts or provide coherent explanations for things like math problems.

“It still may not be able to get even the most common facts right, and for the more obscure stuff you’re just rolling the dice and you have no way of verifying it with the system because it will tend to double down and… insist that it’s correct.”

Robins said mammography scanning is a very different kind of AI, bordering on more traditional machine learning, which is relatively well understood now.

However, the broader issue was how personal data would be used.

“Data is what drives these AI systems, it’s what allows them to evolve, it’s what allows them to be built in the first place. So any data that these AI manufacturers can get, if they can, they’re probably going to want to feed that back into their systems,” he said.

“There are ways to use these systems that are relatively data-safe—you can run local versions of AI, you can make sure that the data isn’t being used to tune things in different ways. So it’s not an insurmountable problem, but it’s a problem that needs to be addressed head-on and addressed in the design of how these things are used.”

The European Union has regulated AI use in high-risk sectors, including education and healthcare, but New Zealand has yet to follow suit. Robins said some sort of guidance was needed.

“Whether that’s legislation or direct regulation or just a relatively high-level set of expectations that the government is to publish, but I think at this point it’s good to have some form of guidance and I imagine that some form of regulation will be inevitable.”