Bridging Gaps in Dental Care through Artificial Intelligence
“The future is human-centered, patient-first, data- and evidence-driven, and AI-empowered.” With this vision, Dr. Natalia Chalmers, Chief Dental Officer and Head of Clinical Innovation at Overjet, opened the second lecture of the Dean’s Distinguished Lecture Series, Building Bridges for Oral Health Research at Columbia University College of Dental Medicine (CDM). In her talk, “Artificial Intelligence in Dentistry: Policy, Evidence, and the Future of Dental Education and Practice,” she explored how AI can improve access, support earlier diagnosis, strengthen consistency across providers, and deepen connections between oral health and overall health. The event also reflects CDM’s growing focus on AI, including a new partnership with Overjet aimed at advancing collaboration in AI, dental education, and oral healthcare innovation.
Chalmers’ perspective is shaped by more than two decades of experience spanning clinical practice, research, and federal policy. A board-certified pediatric dentist with a PhD in oral microbiology, her work has focused on improving outcomes for children and other vulnerable populations. She previously served as the first Chief Dental Officer in the Office of the Administrator at the Centers for Medicare & Medicaid Services (CMS), where she led initiatives to expand dental coverage and implement new quality measures. This included advancing Medicare policy changes allowing coverage of certain dental services for medically compromised patients, including cases where untreated periodontal disease can increase the risk of complications such as pneumonia. With a background in implementation science and translating evidence into practice, Chalmers is uniquely positioned to help guide the integration of AI into clinical care.
Challenges in Integration, Cost, and Access
Chalmers outlined structural challenges across the healthcare system, explaining that medical and dental care often operate separately, with limited communication and disconnected health records creating gaps in care. This can leave dentists without full visibility into a patient’s medical history and drive patients to seek emergency care for dental conditions that could have been treated earlier. At many institutions, these divisions remain deeply embedded in how care is delivered.
CDM’s integrated medical-dental model, however, offers a different approach to connecting oral health and overall health. “As Dr. Chalmers noted, CDM is one of only a few dental schools nationwide to use an integrated medical-dental electronic health records system, Epic,” said Dr. Rena D’Souza, visiting professor and lead organizer of the Dean’s Distinguished Lecture Series. “I believe that the school is best positioned to lead the way in the responsible use of AI and data science for all levels of patient-centered oral health care. This is an exciting time as novel discoveries can be made to firmly establish the bidirectional connections between oral health and systemic health.”
At the same time, fragmentation across the healthcare system is closely connected to ongoing challenges around cost and access. Dental care ranks among the highest areas of healthcare spending in the United States, with approximately $140 billion spent annually, yet outcomes have not kept pace. Much of that cost is paid out of pocket, creating barriers to care, and only about 37% of Americans receive both medical and dental care each year. As a result, many patients, particularly those with public insurance, are more likely to see a physician than a dentist, leaving oral disease untreated until it worsens. While children enrolled in Medicaid are entitled to comprehensive dental benefits, only about 45% receive an oral evaluation nationally. For adults, coverage is more limited and varies widely by state, often extending only to emergency procedures.
Recent policy shifts have begun to expand access to dental care. CMS now allows Medicare to cover certain dental services for medically compromised patients, including those undergoing cancer treatment or organ transplantation, affecting an estimated 1.3 million people. Medicare also offers incentives for primary care providers to incorporate oral health into patient care. Even so, meaningful change has been slow to reach many patients.
The Role of AI in Addressing Gaps in Care
“When we talk about the role of AI, I hope you consider how it will solve the access problem and improve patients’ understanding of their disease,” Chalmers said. She pointed to how these technologies are already improving diagnostic consistency, patient communication, and clinical efficiency.
One of the primary challenges AI can address is variation in diagnosis. Clinicians often disagree when interpreting the same radiographs, and in some studies, only 22% of treatment recommendations are unanimous. General dentists may miss up to 60% of lesions, particularly in early-stage disease. Chalmers attributed this variability to differences in training, image quality, and other clinical factors. AI can help reduce this variation by bringing greater consistency to image interpretation, flagging image quality issues, and detecting findings that may otherwise be missed.
AI systems are developed using curated clinical datasets rather than open web data. Dentists annotate radiographic images by labeling features such as enamel, bone, restorations, and caries, allowing models to learn patterns associated with disease. When applied to new images, these systems can highlight areas that may represent small cavities or bone loss, providing a more consistent reference point for interpretation. Before entering clinical practice, the tools must undergo rigorous regulatory review and demonstrate improved performance alongside clinicians to receive FDA clearance.
Chalmers also emphasized that AI extends beyond diagnosis into how care is measured and understood. Systems can generate an oral health score from 0 to 100 based on overall disease burden, providing a standardized way to assess and track oral health across patients.
These capabilities are also beginning to shape both clinical practice and education. In studies with dental students, AI improved the detection of early-stage disease, particularly among more advanced trainees, suggesting a role in strengthening clinical training rather than replacing it. Students also reported that the technology made imaging easier to interpret and improved patient communication, while underscoring the need for better integration into dental curricula.
In practice, these tools can also reduce the time required for routine tasks. Periodontal assessment can be shortened from about 10 minutes to 2, while charting can be reduced from roughly 20 minutes to 4, easing administrative burden and allowing clinicians to focus more on patient care. Visual overlays also help patients better understand where disease is present, particularly in pediatric settings where parents can more clearly see their child’s condition.
The Limits of AI in Clinical Care
Chalmers was equally clear about what AI cannot do. “It will never replace a clinician, can never perform an exam, never diagnose,” she said. “I want to say this a million times, [it] will never create a treatment plan.” While AI can offer suggestions or highlight findings, she emphasized that care must remain patient-specific and compassionate.
She also underscored that the technology depends on the quality of the underlying data. “It will never replace the need for a good X-ray,” she said, noting that non-diagnostic imaging remains a persistent challenge. In this context, AI functions as a support tool rather than a substitute for clinical judgment, reinforcing the importance of keeping the “human in the loop” at every stage of care.