As futuristic as AI and machine learning are for the dental lab, there is still more room to develop and grow. Expect the technology to become even more adaptive and even more popular, Aguirre says.
“Digital technology and artificial intelligence will continue to improve and grow in popularity as their capabilities and functions expand,” she says. “Both technologies will include more services that benefit laboratory owners, allowing them to save both time and money and also improving production quality and consistency.”
Marrano also anticipates continued growth and understanding of technicians’ workflows. He expects a workflow, similar to Glidewell’s, that will be adopted on a grander scale by the industry.
“Best case for all this is to be able to take our smartest, our brightest, for it to learn their vision of what the output should be, and then be able to replicate that every time in a mass-customization situation,” he says. “Because that’s where it comes in. Everything’s exactly the same, and everything’s exactly different at the same exact time. This is a strange profession. It’s customized mass production. It has to learn from our smartest and our brightest and apply the output that they would, basically, envision as ideal, but apply it to an infinite number of different possibilities.”
The goal of the lab industry needs to change, he adds. Instead of hunting for more digital lab technicians, consider implementing AI software, which has the ability to learn from the brightest lab techs in the industry right now. With AI, a lab designer can go from 50 units per day to 200 units because they would only be approving designs instead of creating them.
“That’s really where I’d like to go with it,” Marrano adds. “That’s my wish list; that’s the future.”
The lab doesn’t just realize the benefits of AI and machine learning. The capabilities and rewards afforded by this technology flow downstream to doctors and, ultimately, to patients.
“I do see a number of directions where it can generate additional breakthroughs, for doctors, patients, and the lab,” Azernikov says. “If you look at our lab, for example, there’s still a lot of manual processes in place. And it’s still not optimized fully, for the tools that are available today. It’s because the labs that you have today were built with tools that were available 20, 30, 50 years ago. Our company existed 50 years, and now we have a new generation, highly automated manufacturing facility, which is built completely from scratch, without thinking about legacy. There’s really a mindset of how we can do it in the best way.”
Tanz predicts even more capability of the machines to keep learning and improving.
“In order for us to level up, fundamentally, and for machine learning to move to an unsupervised learning model, which will be, arguably, much easier to train and much more robust and could take in unstructured data sets and utilize them more easily, there are fundamental breakthroughs that will need to occur from a computer science and research perspective,” he says. “It’s very difficult to predict when those types of breakthroughs are going to occur. It could be tomorrow, it could be in 8 years from now, and it could be longer. So, there are certain fundamental breakthroughs that lots of smart people are working on, and that always has the ability to change the game. So, barring something like that, which obviously is very difficult to predict, I would say that there is a huge rabbit hole that you start to go down.
“I think over the next 5 years and longer, what we’re going to see in the dental category is really the implementation of these technologies in those ways, which are, many times operating behind the scenes and aren’t going to be groundbreaking news,” he continues. “But when you add up all of the efficiencies that are gained in cost savings, and most importantly, the increased standard of care, then I think what you have on your hands is a very profound—maybe the most profound in a very long time—impact to the dental category. No other technology that I’m aware of has the ability to deliver to patients increased care in this manner. You’re talking about better diagnoses, better treatment planning, better restorations, and better review of the quality of the work.”
“Machine learning is here,” Azzara says. “The timeline is now. When we reach that point of mass adoption and mass leverage, to the point where it is a regular part of our day-to-day life, you can equate it to the innovation and adoption of the World Wide Web—it is such a part of our everyday life now that it is seamless. That will happen. Now, it is subtle and hard to say when a complete transformation will take place, but even more difficult to say that it is not here now.”