Less typing, more treating: push for mandatory AI medical skills

Sunday, 15 March 2026

AI tools can ease national healthcare staffing crisis, slashing paperwork and allowing doctors more time with patients, says a Counties Manukau Health medical specialist and digital health expert who uses it on his patient rounds.

Dr Will Reedy, a locum doctor in general surgery, orthopaedics and plastic surgery, says AI tools can free clinicians from hours of handwritten notes and manual typing.

Reedy is passionate about re-imagining health and wellness in the digital age to transform health services. He has worked in Australia, Asia, North America, the UK and Europe.

“AI adoption at scale can help fix the staffing crisis,” says Reedy, health lead at Accenture Aotearoa, which helps companies become more data-driven. He is part of its Global Health team.

“(AI can help) not by replacing people, but by freeing them from administration so they can get back to healing.”

Technology could also help address what he calls the health system’s “postcode lottery”, where people in wealthier areas tend to receive better care.

“The number one data item that determines your health and wellness is your postcode,” Reedy says.

“If you have a postcode in Whanganui, generally speaking, your health outcomes are worse than living in Auckland CBD. The health system, literally, is a postcode lottery.”

Reedy argues AI use should be mandatory.

“Like the stethoscope around my neck, this is a clinical tool,” he says. “I think about data and AI as a human and machine partnership.”

“Not by replacing people, but by freeing them from administration so they can get back to healing.”

But adopting it will require a major shift in how clinicians work.

“How do you get the workforce to work in this human-and-machine model?” he asks.

“Do you start mandating it in medical schools? You actually need the whole current workforce to upskill.”

He says Health New Zealand, a $32 billion organisation, should consider investing in a national training programme.

“You could create an academy to train all clinicians in AI tools.”

Health Minister Simeon Brown last month announced every emergency department had access to an artificial intelligence scribe tool “Heidi”.

Medical “scribes” listen to a doctor–patient conversation then produce clinical notes, paperwork, and summaries that take doctors a lot of time to work up.

An initial pilot highlighted their practical impact. Doctors using the tool were able to see, on average, an additional patient per shift as a result of time saved.

“That means faster care for patients and shorter waiting times in our emergency departments, while also reducing pressure and cognitive load on busy clinicians,” Brown says.

After one month at Middlemore Emergency Department, 80% of surveyed staff said the AI scribe improved productivity or efficiency.

More than eight out of ten said it had a positive impact on their overall experience and wellbeing during a shift.

Health New Zealand is now chasing more than 1000 additional licences, predominantly for use by mental health teams.

Reedy was already using AI during his hospital ward rounds. He records voice notes about patients, then uses AI software to generate clinical summaries.

The system saves about 20 minutes of paperwork per patient, he says.

But he stresses the technology must still be checked by clinicians, as one danger is medicines can have similar-sounding names but massively different uses.

“Medicine is a knowledge industry, and we rely on the data. But the art of medicine is when I look at that data and ask: ‘Do I think that’s actually what’s happening with this patient?’”

Reedy hopes within a year he will no longer need to type notes during ward rounds.

“I would love not to typing or writing anything, just to talk and a really high-quality clinical note appears.”

Significant barriers remain.

New Zealand’s healthcare data is scattered across more than 6700 organisations, from GP clinics and pharmacies to aged-care providers and hospitals.

Reedy says the data-dependent system needs modern, centralised digital infrastructure.

“There are huge gaps in our data,” he says. “We probably only capture about 20 percent of the information that determines a person’s health.”

Much of the missing information relates to the “social determinants” of health - factors such as housing, income, and lifestyle.

Security concerns also remain. Recent hacks of digital health platforms such as MediMap and Manage My Health have fuelled public scepticism.

Association of Salaried Medical Specialists (ASMS) executive director Sarah Dalton says those breaches served as warnings that digital tools and the records they hold had to be secure.

Dalton says ASMS is closely following the uptake of the AI medical scribe Heidi, being used by members in emergency departments across the motu.

“While overall this is a good thing, we do need assurances about where the information will be stored – and that patient privacy is maintained,” she says.

“Digital sovereignty is a key concern – we need to keep patient information here in Aotearoa.

“We also want to ensure AI scribes are never used as a substitute for real human staff in hospitals or healthcare settings as a form of health workforce rationing.

“As it stands, we are so far away from a fully-staffed health system that we need all the tools we can get.”

And while some do worry AI could replace healthcare jobs, for now Reedy believes the technology will mainly replace repetitive tasks such as paperwork.

“AI won’t replace the gut instinct of doctors and surgeons,” he says.

“But it can replace the pen and paper.”

He predicts healthcare will eventually move beyond today’s generative AI systems, which produce text, towards “agentic” AI tools capable of carrying out tasks on behalf of clinicians.

DCE Tech and Data Michael Dreyer recently outlined developments in AI and tech aimed at reducing health sector administrative burdens, stabilise payroll systems, and improve clinical decision-making.

The Clinical Timeline tool, which consolidates and summarises data about an individual patient, is expected to save clinical staff up to 60% of time on administrative tasks, freeing them up to do ‘”high value” work, Dreyer said.

It is in the test phase with some features expected to be rolled out in March and April.

As a word of warning, a 2026 Kordia New Zealand Business Cyber Security Report of nearly 250 businesses with 50-plus employees found staff using AI improperly was one of the biggest cyber security challenges.

Cyber-attacks carried out through AI vulnerabilities more than doubled, from 6% in 2024 to 14% in 2025.

Patrick Sharp, general manager of Kordia-owned Aura Information Security, says it was often down to vulnerabilities caused by businesses implementing AI systems without sufficient consideration of security.

“Insider threats, whether accidental or malicious, have always been a factor in cyber incidents and data breaches,” says Sharp.

“But shadow AI - the unauthorised use of AI tools by employees - is growing into a massive problem.”

Individual staff were copying confidential data into AI systems without understanding the risks and without guidance from their organisation, many of whom were implementing sanctioned AI tools without sufficient security governance.

Sharp says there is a real part for Government to play.

“New Zealand’s cyber security legislation lags far behind our global peers,” says Sharp.

“How prepared are we, and are we investing enough into our collective cyber defence?”

Most attackers don’t hack in, they log on. Securing identity is key, Sharp says.

Controls, such as phishing-resistant multi-factor authentication (MFA), least-privilege access, continuous verification, and robust password reset processes, can significantly reduce risk.

Organisations need a combination of layered security controls, validation of their effectiveness, appropriate monitoring, and practised incident response.

Privacy Commissioner Michael Webster says the Privacy Act requires organisations to think about how they are protecting information when they use AI tools.

“Our 2025 privacy survey shows many people have concerns about AI, with 62% saying they were concerned or very concerned about business or government agencies using AI to make decisions about them,” he says.

“Thinking about privacy is vital to using AI tools well. We recommend conducting a Privacy Impact Assessment when a new technology or operating system is used,” he says.

A PIA helps agencies identify and assess the privacy risks arising from their collection, use or handling of personal information. It will also propose ways to mitigate or minimise these risks.

Patients needed to be informed about the use of AI and organisations need to consider giving a fair chance for individuals to opt out.

The promise of artificial intelligence in healthcare, Reedy says, is not about machines taking over medicine.

It is about giving clinicians back the one thing the modern system has slowly taken away: time.

“The best part of medicine is talking to patients,” he says. “If AI can take the typing away, that means more listening, more thinking, and better care.”