Dr. Peter Fritz, a certified periodontist and lifelong learner, discusses artificial intelligence and it’s place in dentistry – how it is being used now and what the future may hold.
Read the audio transcript below:
Dr. Luisa Schuldt (LS): Hi everyone. Welcome to Brush Up, presented by Oral Health Group, the dental podcast where we speak with industry experts to discuss a variety of topics such as technology, finance and practice management. I’m your host Dr. Luisa Schuldt, a prosthodontist and periodontist based out of Fonthill, Ontario.
Today, I have a very unique pleasure. I have the opportunity to present someone I work with pretty closely and I get to see almost every day. Also based out of Periodontal Wellness and Implant Surgery in Font Hill, Ontario, today, we’re joined by Dr. Peter Fritz to talk about artificial intelligence and its place in dentistry. Dr. Peter Fritz is an intrepid, lifelong learner, certified as a periodontist, and is in full time private practice where he gets to lead a collaborative, empowered team of clinicians, myself included. We are a group of scientists, explorers, and artists who all perform together to innovate the dental specialty of periodontics and redefine the patient experience. As a futurist and advocate of perpetual change, Peter sets a clear and ambitious path to success based on scientific rigor, inspiring others, and clinical excellence. He strives for improvement, guided by his lifelong goal in all elements of life. Give me success or its eternal pursuit and I’ll take the pursuit. Just a little bit more background on Peter. He also has an MBA and a law degree, which also gives him a very unique perspective into the matter on hand today, which will be AI, artificial intelligence in dentistry. Peter, welcome. Thank you so much for joining us.
Dr. Peter Fritz (PF): Thank you, Luisa. It’s a pleasure to be here. Thank you to Oral Health for inviting me to speak. And it’s a pleasure to have this upcoming discussion with you.
LS: Well, we get to discuss many things on a day-to-day basis. Today, I’d love to focus on AI. And if you don’t mind, I would love for our listeners to start by just learning a little bit more about you and how this journey in the world of AI started for you. Where did this interest in artificial intelligence start from?
PF: Well, I I belong to this group called the World Innovation Network where we get together in Chicago every year and discuss what the future is of different sectors and about 10 years ago, they were starting to talk about how AI is going to transform every sector and we should just keep it on our radar. Well, you know, that discussion just got stronger and stronger like a building wave. And in December of 2022, I was on a on a call and ChatGPT was just released, and they said you have to look at this, this is going to change everything. And sure enough it became real for me at that point, and since then we’ve been using it almost every day in our practice and it’s just a phenomenal transformation of what is to come.
LS: To give our listeners a little bit more background on Peter and his journey in AI. He had the opportunity to lecture in a few different forums on this topic, including with The Royal College of Dentists of Canada, the RCDC, and you have a couple more lectures coming up based on all that knowledge you already have about AI. How is it being used currently in dentistry? What can we do with?
PF: When you’re stepping back a little bit, when you think about how it works, it’s kind of like you have these ingredients. That you’re going to a restaurant, you have these ingredients that you bring with you and that’s sort of the prompt. That’s the information that you give the program, then it goes into the kitchen and with some magical algorithm or some magic that happens in the kitchen, which is hard to describe other than it’s there’s a lot of sophisticated algorithmic computing that goes into that. It takes these ingredients and processes it into something. In an instant, the meal is brought to you and they say see what we’ve created. This is what we have and then all of a sudden, you’re like, wow, this is amazing. I can’t believe they did this. And occasionally it’s like, oh, geez, I can’t believe they did this because it’s not exactly what we’re expecting. So, it all depends on the ingredients that go into the algorithm to make the output come through. But AI, it’s going to change things for dentistry, much like the internet did. I remember in 1997, I graduated dentistry, and I didn’t have an e-mail account and we had no Internet at that point or e-mail. Can you imagine practising today without the Internet? Well, when we have this discussion in five years again, you’re going to find that practising without AI in your practice is just going to be as foreign as not having the internet. So, it’s kind of like an AI program for dentistry is like having like an assistant there for you, who amplifies your clinical abilities and administrative tasks. So, in diagnostic imaging, for example, it can detect pathology like caries and periodontal disease and abscesses just by looking at the radiograph and comparing that to a bank of millions of other radiographs. With consistency and precision that is just getting better and better. So, it’s not like fantastic right now. I think like a year ago it was like a C; now it’s probably like a B minus and B maybe you know. But in half a year, I think it’s going to be just developed that much more. And things like workflows, administrative workflows, like how you create a report or how you do your charting, this is all affected by AI. And it frees up clinicians to focus more on patients so you’re able to focus more on direct patient care. So, we’re having different tasks that we’ve sourced out and allowing us to have our powers. Our superpower is augmented. So,now that we’re going to be stronger clinicians than we were before, it’s not going to change the need for clinicians. It’s just going to be a much stronger thing. So, two things are certain. AI is here to stay, and no one is an expert.
LS: Yet! Not even AI. Yeah, I really like the way you put that, that AI, it is learning as we are using it and it’s just getting better and better and the processes are getting smoother as we use them more and more often. You mentioned already one of the specific current applications, radiography and imaging. It’s also used in labs for design. Our digital impressions are reinforced and powered by AI as well. Do you see any other specific applications that are revolutionizing the way we practice today?
PF: Well, one of the big features of these convolutional neural networks, or CNN’s, is that they’re able to recognize patterns so well. And what dentists do really well is we also recognize patterns in X-rays and patterns in histology slides and those things. So, it’s designed to detect anomalies in radiographs and CBCTs. And, you know, companies like Prototype have been using this for looking at histological diagnosis in terms of predicting cancers as well. So, radiography is really the biggest, the most exciting part, that’s most developed at this point, I think. And there are a number of different companies like Overjet, Diagnocat, Pearl, Deep Care. These are all up and coming. I think they’re going to get more and more amplified as time goes on. AI is also integral to orthodontic treatment, because now you can sort of predict how to simulate the tooth movements and predict how the outcome is going to be for aligner therapy. It can help you place implants so you can do virtual surgeries that are guided with AI to help us determine where the structures, the critical structures are. So, it improves safety and decreases treatment time and as well the administrative tasks. I can’t stress about how time saving that is for people that, you know, for writing your chart notes and for writing letters and even things like performance reviews and letters to the insurers are clearer and much faster than before. For those that have had to write a letter to describe the clinical condition to an insurance company to kind of describe things better, you know, you can do this now with just a really good prompt and it can really get the message across, and you have most of the information there. It’s not to say you shouldn’t edit it because you always need to double check what it does, but at least it gives you some structure that can save you a ton of time.
LS: You were mentioning front desk and administrative functions as well. Am I understanding correctly that there are AI systems that can, you know, pick up the phone and navigate a phone call to a certain degree of, you know, clarity and answer patient questions or concerns?
PF: There sure are. Like with the chat bots that you know on websites, these computerized chat bots which are inherently…There’s risk to having a chat bot on your website. So, I would be very careful before you did that, because any sort of patient information has to be protected. We have to be very careful with what we share with the AI to make sure that the AI programs that you’re using are PIPEDA compliant or will be able to be compliant with the new artificial intelligence and dentistry and Data Act, the Data Act which is coming. So yeah, you can do simulations for phone calls. You can actually have it answer a phone call, although we’re not at that point yet. But it’s really not that far from it and I’ll show you in a little bit what I mean in terms of capability with having a personality and being able to express some humor as well.
LS: Now, I love that you’re bringing into our conversation all these things from your background. So, I mentioned earlier to our listeners, your MBA and your law degree and the fact that you are so perceptive to some of the risks in using AI is really good that you’ve mentioned. So, this patient information management. Do you see any challenges in the area of, you know, patient consent?
PF: Oh, absolutely. This is, you know, we are just at the very beginning of this and integrating AI into dental clinics isn’t without challenges. First, there’s a huge learning curve for the dentists and the staff members who are, you know, faced with you got to start learning how to use this. So, you know, first that wasn’t part of their employment contract. Now you have to train them to stay relevant with that, they have to adapt to new software and new workflows. They need comprehensive training to maximize the benefits of AI. There’s also financial cost to implementing this, which are not insignificant, including the training costs, and this can be expensive. But I think the plus at the other end is that there is going to be lots of cost savings and the skills that you needed before in your team members are going to be different from what they are down the road. So, there’s also interoperability concerns, right? The AI programming not blending very well with your electronic records system or your imaging equipment, so you may have to have additional IT support to bridge these platforms together and that’s frustrating. That never goes fast enough. And then as you mentioned, patient security, patient data security is a huge concern. You have to have a robust cybersecurity measure. You know, if your practice doesn’t have cyber security insurance right now, you should stop and after this podcast, go contact your insurer because you’re liable or you’re vulnerable if you don’t have this insurance for cybersecurity attacks. And they’re also ethical implications. Like, what about the biases in the data? We have to be concerned with how, what the bias is, is it is it saying that these caries that the AI radiographic software finds, is it really aggressive or is it missing things. So these biases need to be considered. Because they can skew your diagnostic outcomes if you don’t address them and it’s just a tool. That’s still your clinical judgment that’s at the end of the day, that makes that decision whether or not you cut the tooth. But the AI can certainly pick up things that you may not see, and in my example, I’ve seen pulp stones more than I’ve ever seen before in my career, because the AI picks up pulp stones and I was not really concerned as a periodontist for checking out where the pulp stones were.
LS: I’m really happy you bring up the fact that the responsibility still lies in the clinician and our decision making as far as the impact of these AI powered tools on treatment planning and decision making. Do you see them taking away from our clinical responsibility from a legal standpoint? I would imagine not, but…
PF: Yeah, we’re shifting the trust from the clinician to the AI. I think that’s the danger that the treatment plan now…So if you’re using some software that can diagnose the cavity and it says it has 70% confidence that you’ve got decay at 2-6 buccal, now it’s almost like you’re hooking, you’re doing a diagnostic on a car, which tells you how much brake life you have left. But we’re not dealing with cars. We’re dealing with humans, and there’s an ethical obligation to make sure that the information that you’re giving your patients is the right information. So, you know, although the information is really handy to have, it doesn’t take the risk away from the dentist at all. It’s just a tool that you would use to make a better data-driven recommendation to provide personalized care. And although AI can, it can like synthesize with this clinical information and it can take the radiographs, patient histories, and it can sort of meld this together to help you, still the clinical judgment and the responsibility rests with you. So, you know, I think it’s really important that we never lose sight of that. However, the power of it helping the treatment plan and giving options and figuring out what the probability of success is for these different options is taking us to a place where we never have been before, which, as a periodontist you know, is really amazing because we have to prognosticate how long teeth are going to last. Now you have this information available. Well, I can tell you what’s the effect of affrication and mobility on tooth 2-6 in a patient with this many other teeth. As a non-smoker that has this genetic polymorphism and you put all those things together and now she should be able to reasonably figure out how long things are going to last.
LS: I guess putting it in our daily clinical context, many of us have very, very talented assistants or hygienists who have been in the world of dentistry for many, many years, and they can look at a radiograph after it’s been taken and they already know what the clinicians are going to say to the patient. But it’s still in the clinician’s hand to make that diagnosis and set up the treatment plan. So AI might detect these things and give us the prompt, but it is still our responsibility to do that diagnosis and treatment planning
PF: Absolutely.
LS: How can AI assist us in personalizing our patient care, especially you and I, we’re specialists, and I think that’s something that’s happening more and more in general dentistry as well. Each treatment is not a cookie cutter type treatment plan, which might be easier for AI to set up. We need things to be still personalized, custom done for the specific needs of each of our patients. Can AI still support us and assist us in that scenario?
PF: Yeah. So, you know, the personalization of AI. First in personalizing patient care is one thing and when people think about AI, they think about Siri and like that is just going to be a straight you ask it a question, it gives you an answer. But the difference with AI is that it’s more of a companion than a crutch, and I think what I’m going to do, I’m going to demonstrate something for you because I’m getting tired and I’m going to let the AI answer your question for you. This is a program called Pi, and I’m going to run it now, so…Hello, Pi, my virtual assistant. I’m doing an interview with Brush Up Podcast with Luisa Schuldt, and she’s just asking me a question. I need your help answering. Is that okay?
Pi: Of course, I’m happy to help you with your interview. What is the question that Luisa Schuldt asked you? I’ll do my best to provide you with a thoughtful and informative answer.
PF: Thanks, Pi. The question is how can AI assist in personalizing patient care and treatment? And I don’t want just a generic answer, I need it with some punch.
Pi: You got it. Here’s an answer that packs a punch. AI is like a dental detective that helps dentists crack the case of each patient’s unique needs. Turning one-size-fits-all treatments into personalized dental experiences that leave patients grinning from ear to ear. In a more detailed answer, AI can personalize patient care by 1. Analyzing patient data and medical history to identify individual risk factors, preferences and needs. 2. Providing personalized treatment plans that are tailored to the patient specific dental issues and goal. 3. Monitoring the progress of treatments and adjusting them as needed based on the patient’s response. 4. Identifying patterns and trends in patient data to predict potential issues and prevent them before they occur.
PF: Thanks, Pi. So, Luisa, isn’t that amazing that it can do that?
LS: That’s like Siri with superpowers.
PF: I agree with the answer that she gave as well. So yeah, so that’s how I think it can help in personalizing it. So, I go with those four points.
LS: Those are really great points, Peter. Thank you, Pi. We were chatting a little bit already about some of the ethical implications and responsibilities using AI and dentistry. Is there anything else you would like to add to the topic of, you know, patient privacy and data security? You mentioned already just how important it is if we don’t have any kind of coverage for cyber security, to talk to our insurers, talk to our AI team as well to make sure everything is well encrypted and protected. Can you sense any other ethical implications?
PF: Yeah, when you’re using AI for patient care, you have to prioritize the data security. You want to know where the data is going, and you want to make sure that that’s kept private, that it’s ethically used, that it’s used for the good of the patient. You’re not trying to trick somebody, you’re not trying to find things out for nefarious reasons, and you want the patient’s consent. To do this, you just can’t force this on patients, we have to establish consent that we’re using AI in their healthcare needs. An example is if for facial recognition, that, you know, if you go into a hospital and they have facial recognition, they can see if that person has been at that hospital emergency room, you know, 50 times in the last month seeking for a drug habit or if they’ve never been there before or if it’s their first time they have a cardiac history and they need to be at the front of the line. The facial recognition can be really powerful. It can also, in addition to facial recognition, it can do emotional recognition so it can tell how stressed the patient is or if they’re in pain or if they’re happy. Which I think is really cool for in the studies where they’ve had dementia patients, and they can tell if they’re really suffering pain based on facial recognition and monitoring of this. But yeah, back to your question, I went on a tangent there. But you need strong encryption and you need to have a place to protect the sensitive data and you want to make sure that the providers of the AI assets that you have are safeguarding that for you. And you also want to make sure that, we mentioned about bias, the bias is actually necessary and it’s inevitable. For example, our operating room is biased towards being sterile and clean, right. But as long as that bias is transparent and it meets the objective, then bias is actually required in a lot of things, including AI, to make sure that we’re meeting the objective. It just has to be known to the people using the equipment. And this creates a lot of complexity and chaos, and you have to respond to complexity and chaos through differentiation because you can’t innovate unless you go through that phase of complexity and chaos. It just doesn’t work. You don’t go into a new space without having that difficult space ahead of you.
LS: For that kind of learning phase, there’s so much to say in so little time. Where do you see this going in the future? Do you see a direct impact from AI in the future? Any upcoming trends with this technology, anything that’s going to significantly impact the way we do dentistry? I mean we’ve already mentioned so many things along those lines and that it’s going to be easier, smoother, hopefully benefiting our ability to customize patient care, make sure patients are cared for well. Do you see any other advancements or trends with AI?
PF: Yeah. So, AI is going to equalize the knowledge that we have access to. So, all of a sudden, a dental assistant has the same information to access or the same information as a dentist or as a specialist does. So, the curation of that AI product or the information has to be really what sets things apart. The AI is not going to replace dentists, it’s only going to replace dentists who don’t use AI. See that’s the problem. And the big trend is going to be regulation, because right now the regulatory standards are in flux. And I’ve been working with some of the colleges to help put that together. And the analogy I give is like, okay, let’s go back 120 years at Kitty Hawk when the Wright Brothers, they took off for a 12 second flight on their airplane, right? 12 seconds up and down. They didn’t have any regulations about flights. When this is about its infancy, nobody really cared about how regulated this is. AI – people are starting to care now. So, in the early days of aviation, there are very few rules like you didn’t have to be licensed pilot and, overtime, the number of aircrafts and the number of flights increased. So did the need for regulation. So, you have the Federal Aviation Association now create standards. Well, we’re gonna have the same standards for AI. And there’s going to be ongoing adaptation because, you know, the regulation of air travel has changed continuously. Look at what happened at 9/11, how we changed going through airports, right? AI similarly is going to be shaken by things that we can’t see right now. And it’ll have to be regulated differently. And coordination like the global travel was not thought about when Kitty Hawk happened, and the same thing with AI. It’s global. It’s happening all over the world. So, our rules are behind maybe what’s happening in Europe about AI. And this technology crosses borders and there have to be global standards. So, I think the pioneering flights that led to the comprehensive and sophisticated regulatory framework about aviation, we’re going to see the same thing with AI. It’s just we’re starting right now; we just start taking that little flight. And we’re starting to think, well, you know, as we fly more, we better get some rules.
And I’m very honored to be part of the task force for the Royal College of Dentists of Canada that is putting together some thoughts about how we approach this and basically how we teach dentists how to use this responsibly and to be compliant with the with PIPEDA because PIPEDA is going to be replaced by Bill C27 in I think about two years and that’s going to change the way that we are our custodians of patients privacy.
LS: Well, thank you so much, Peter, for sharing all this information about AI – where it is today, where it could go potentially in the future, how it can make our lives and our patients lives better, richer. Also, Peter, I forgot to thank you for inviting another guest today, Pi, to join us because that was a really nice unique piece. Thank you as well to our listeners. Thank you for listening today. Be sure to subscribe on Spotify and follow us on social media to be notified every time we post a new episode to our listeners. Keep brushing up!