Dentist Sen Le says he is not afraid of losing his job to AI: a technology he has helped to develop, and one which he says can detect problems in a patient’s mouth that most human dentists would miss.
asserted
dentists → say → that
“The human eye can only see about 40 shades of grey, but the AI can see about 64,000,” he says.
asserted
he → see → 64,000
“That means it’s able to see things that we can’t: subtle changes in anatomy and subtle changes in disease.
asserted
we → mean → disease
Often, it will find something that I didn’t even notice was there.”
asserted
I → find → something
While there is debate about the extent to which AI could destroy human jobs or create new ones, the technology is already well documented to be changing the nature of roles in various industries including administration, law and media.
uncertain
technology → be → administration
But it is also being used – and developed – by people in more hands-on professions ranging from dentistry to run coaching and restaurant management, some of whom say AI is improving the services they offer to customers but will never replace them.
asserted
they → use → them
Giving diagnostics teeth
Le started working on software called Eyes of AI six years ago to save dentists time and help them make more consistent and accurate diagnoses.
asserted
them → give → diagnoses
“Dentists and dental professionals are very time poor,” he says.
asserted
he → say → ?
“There’s also a lot of variability in our diagnostics: if you go to five different dentists, you get five different diagnoses, and then the treatments they recommend could range from $300 up to $36,000.”
uncertain
they → ’ → 36,000
The Sydney-based dentist, who also works in Canberra, has used the technology over the past three years with his patients and says it can identify issues ranging from defective fillings to infected abscesses and calcification – the build-up of excess calcium – in places not just isolated to the teeth.
asserted
it → base → teeth
The AI tool can work with as little information as a single X-ray to make basic findings, but can also consider other clinical information fed into it.
asserted
tool → work → it
“It’ll detect issues such as tooth decay early so that we can address it very conservatively rather than drilling into the patient,” he says.
asserted
he → detect → patient
“I also use AI as a second set of eyes – like a spell check – to make sure that I haven’t missed anything, especially if I’m busy or running late.”
asserted
I → use → anything
Le and his team have been partnered with the CSIRO and the University of Sydney, and while tighter regulation has made the approval process less straightforward, he says the business is on the verge of getting the technology approved for release to other clinics around Australia.
asserted
technology → partner → Australia
AI can also enhance patients’ understanding, Le says, helping them to take more ownership over their treatments.
asserted
them → enhance → treatments
“Often, patients are just taking the dentist’s word for what’s there,” he says.
asserted
he → take → word
“But the colour overlays generated by the AI, and the things it flags, can help patients look after their oral health better.”
asserted
patients → generate → health
While Le believes AI will increasingly play a supporting role in dentistry in years to come, he says it will not dampen demand for dentists.
asserted
it → believe → dentists
“It’s a very human thing to go into someone’s mouth,” he says.
asserted
he → ’ → mouth
“And at the end of the day, AI requires human oversight because it can’t be held responsible for decisions.
asserted
it → require → decisions
I don’t believe it’s possible for AI to replace dentists.”
asserted
AI → believe → dentists
Running the numbers
Sydney-based trail and ultra-running coach Ash Daniels started experimenting with AI two years ago.
asserted
Daniels → run → AI
“It was like a new employee at first,” he says.
asserted
he → say → employee
“I had to check the output, but I gave it more and more tasks and built trust over time.
asserted
I → have → time
Now, it’s like having an assistant.”
asserted
it → ’ → assistant
While apps are increasingly being used to generate training plans for athletes, Daniels says some of them can lead to people overtraining and getting injured.
asserted
people → use → them
“What they lack is a feedback loop,” he says.
asserted
he → lack → What
AI can feed back some data, but it lacks empathy, and if you’ve had a real tough week at work, or your kids haven’t slept, AI can’t adapt to that.”
asserted
AI → feed → that
Daniels says he is not worried about AI taking his job, and that he has so far seen no drop-off in the number of people he coaches.
asserted
he → say → people
“Huge amounts of people are coming into the sport and looking for guidance,” he says.
asserted
he → come → guidance
“They want to speak to me, not a machine.”
asserted
They → want → me
AI is useful for doing repetitive administrative work, Daniels says, as well as analysing, identifying patterns, and concisely presenting findings from huge amounts of data.
asserted
Daniels → do → data
“Before, I would have to trawl my way through data files from running watches and other digital devices,” he says.
asserted
he → have → watches
“Now, I can do that more quickly, and AI will pump it out into a digestible PDF format people can easily understand rather than a spreadsheet with a load of data.
asserted
people → do → data
It also means I’m more available to chat through things with athletes.”
asserted
I → mean → athletes
But Daniels says coaching is about more than just generating a plan, or analysing data, and that things such as lived experience and shared moments cannot be substituted by AI.
asserted
things → say → AI
“My mantra is that a plan will get you to the start line, but I’ll be waiting for you at the finish line,” he says.
asserted
I → get → line
“You’re never going to get that from a machine.”
asserted
You → go → machine
Serving up
AJ Eserjose, co-owner of Charcoal & Crisp in the Sydney suburb of Crows Nest, says AI is helping businesses like his save time on tasks such as rostering and preparing marketing.
asserted
businesses → serve → marketing
“Most of our time is consumed by repetitive admin work,” he says.
asserted
he → consume → work
…and 11 more, not listed.
The Aporia
