Will AI replace dentists?

The topic of whether robots will ever replace humans at the workplace has dominated discussion panels and forums across the globe.

Obviously, artificial intelligence is playing an increasingly important role in the medical and dental industry. There are good reasons why ‘autonomous technology’, ‘robotics’ and ‘artificial intelligence’ are also buzzwords in the healthcare industry.    

From intelligent diagnosis to intelligent treatment, smart health management to dental surgical robots, it seems there’s no escaping the looming subject. If you’re feeling a bit nervous about where this is all heading, you’re not alone.

ChatGPT and the Future of Dentistry

If tech visionaries like Bill Gates are warning that computer-chipped intelligence are coming for our jobs – or in Elon Musk’s case, destroy the universe, whichever comes first – perhaps we will do well to give the topic the attention it deserves. 

Ultimately, the question on the lips of dental professionals is this: will doctors be replaced by technology?

No mean feat

We’re living in an age where computers are able to process a ton more data than the average human, and for the first time, deploy machine-learning algorithms to make “sense” of them.

Take for instance, AI applications in dental implants. They include: preoperative digital 3D scanning of the implant site; digital implant surgery design; design and manufacture of surgical guides; and real-time navigation during surgery. The latter has been proven to improve the accuracy of artificial dental implant surgery, reduce surgical trauma, and shorten surgery time.

These are by no means repetitive or menial tasks that you would normally associate with a robot. They have obviously come a long way from sorting boxes at an Amazon fulfilment centre!

World’s first autonomous dental implant surgery robot

On 14 January, 2021, the world’s first autonomous dental implant surgery robot developed by Prof Zhao Yimin and his team from the Chinese Academy of Engineering in 2017, successfully performed dental implant surgery on the 100th patient. This momentous procedure was carried out at the Digital Oral Medicine Center of the Stomatological Hospital of the Air Force Military Medical University.

The robot not only accurately completed the implant placement of a missing tooth according to “pre-determined instructions”, the whole operation only took 10 minutes.

Professor Zhao Yimin was able to place the pre-designed 3D printed instant denture on the implant, restoring the integrity and occlusal relationship of the patient’s dentition during the same appointment. Patients’ feedback of their experience receiving treatment from the “robotic surgeon” was “easy and comfortable – no fear at all”.

Robots are not just into dental implants either.

They have been successfully applied in oral and maxillofacial surgery for tumour resection in the OMS region.

Some of their complex applications include: Preoperative oral and maxillofacial 3D image data acquisition and reconstruction; accurate analysis of lesion characteristics; production of guide plates to guide surgical operations; accurate segmentation, remodelling; displacement and fixed.

Dentists lost to AI system in predicting tooth decay

In the US, an AI system beat dentists at diagnosing tooth decays based on bitewing and periapical radiographs.

The study was commissioned by Pearl, an AI start-up that deploys innovations in computer vision and deep learning technology to help dental practices maximize patient care and operational efficiency.

According to the company, the AI system was pitted against three experienced dentists in analyzing 8,767 bitewing and periapical radiographs.

Not only did the study show the AI system superior both in terms of accuracy and consistency in its tooth decay predictions, it revealed a lack of diagnostic consensus between the clinicians. They only agreed on the presence of decay in 4.2% (370) of the x-rays, although they were in agreement on the absence of decay more often – 79% of the time.

While these were experienced dentists, the machine learning system had an “unfair” advantage, having studied millions of x-rays evaluated by hundreds of human dentists!

Pearl is not only helping out dentists but dental lab technicians as well.

Its Scan Clarity Score app allows labs to instantly determine the quality of intraoral scans at the prepped tooth level.

The AI scan processing software is able to quickly identify problematic issues such as lack of a visible margin or defects in the scan. The smart system then grades each intraoral scan based on margin clarity. Should the score be too low, it would be flagged for a lab technician to intervene.

Self-oral diagnosis – using a smartphone

In the pandemic era for dental emergency care, Artificial Intelligence (AI) can also help patients self-diagnose their own oral problems using a smartphone. 

For example, an AI–powered mobile phone app known as AICaries can be used by parents and caregivers for dental caries detection.

Using a two-step qualitative study design, the app provides interactive caries risk assessment and personalized education on reducing children’s Early Childhood Caries (ECC) risk.

Using their smartphones, parents can take photos of their children’s teeth and detect ECC aided by AICaries.  This helps parents to actively seek treatment for their children at an early stage of ECC.

Not so fast, buddy!

By identifying caries and other dental pathologies earlier and faster – or as shown, more accurately than a dentist – these AI systems provide a valuable “second opinion” that will help practitioners avoid diagnostic mistakes.

In so doing, they allow dentists to focus on doing what only humans can do, which is to communicate treatment plans with compassion and empathy.

No matter how sophisticated AI technology becomes, robots can only rely on existing data for diagnosis and treatment. However, even the same type of disease will have different symptoms in different people.

In the face of symptoms that have not appeared before, robots have no ability to judge and solve them. We will still need to rely on a human doctor to administer the oral examination and treatment.

Whether it is an dental surgical robot or a robot for artificial intelligence diagnosis, the underlying foundation still depends on the doctor. The more sophisticated and complex robots are, the more doctors need to learn, master and operate them.

The artificial intelligence of medical diagnosis also requires doctors to accumulate the cases they have handled, and then turn these into data and store them in the artificial intelligence database, so that diagnosis and treatment judgments can be made based on similar conditions.

While educational and informative apps can help to instil self-awareness and allow patients to have better control over their oral health status, there is no substitute for a human dentist exercising a sense of timing and sensitivity to better understand the patient’s needs, and tailoring a treatment plan accordingly. 

At the speed of AI

On the other hand, robots and artificial intelligence have been advancing at warped speeds beyond anyone’s imagination.

In the case of robots performing precise dental implant placements, the AI system needed to realise many mind-blowing algorithmic calculations: Interpretation and reproduction of the anatomical structure of the surgical area; the precise planning of the implantation plan; the real-time navigation and calibration of the operation; and the automatic and precise implementation of the whole implantation operation.

To think that dental implants are still considered a relatively modern technology. This is an area that not too long ago was heralded as a massive breakthrough. How soon AI would be able to take over the operational delivery on scale is anybody’s guess.

By comprehensively applying a wide range of applications – including machine vision, mechanical sensing and micro-modular robots – AI-enabled systems can solve many problems once thought impossible, or at least once considered a long way off. In the last 3-5 years alone, AI has improved by leaps and bounds in the areas of spatial mapping, precise alignment, and follow-up control.

What these AI solutions have shown are clear advantages of predictability, reproducibility, precision and efficiency – in addition to being minimally invasive and safe.

The more these so-called roadblocks to a fully realised AI future are being removed, the closer humanity moves towards a technologically bright yet to some, psychologically unnerving future. The sci-fi inspired conundrum of who wins in the “man versus machine” debate may also finally reach a conclusion.