La Presse Investigation: AI in Dentistry Raises Overdiagnosis Concerns

Artificial intelligence is increasingly integrated into dental diagnostic workflows, with recent investigations revealing that AI-assisted screenings can significantly alter treatment plans compared to traditional human-only assessments.

As clinics adopt these software tools, patients face new challenges in dealing with conflicting diagnostic outcomes and the financial implications of potentially increased treatment recommendations.

Clinical Insights for Dental Patients

  • Diagnostic Variance: Studies indicate that AI may identify more potential lesions than human clinicians, leading to a wider range of recommended procedures.
  • Incentive Alignment: Software manufacturers often market these tools based on their ability to increase patient treatment acceptance rates, which may influence clinical decision-making.
  • Critical Verification: Patients should feel empowered to discuss diagnostic discrepancies with their dentist, especially when presented with high-cost treatment plans based on imaging software.

The integration of artificial intelligence in dentistry has moved from theoretical research to active practice, creating a complex environment for both clinicians and patients. An investigation by La Presse journalists Gabrielle Duchaine and Ariane Lacoursière highlighted this tension by comparing diagnostic outcomes in the same patient, according to Lapresse.

While dentists relying solely on traditional visual and tactile examinations identified between zero and three cavities, the integration of AI software suggested up to seven lesions, including one tooth requiring extraction.

This discrepancy mirrors the frustration often associated with home improvement projects, such as installing a heat pump, where conflicting expert opinions regarding technical requirements and costs leave the consumer without a clear path forward. The core medical concern is the risk of overdiagnosis—a scenario where a diagnostic tool identifies minor anomalies that may not require immediate clinical intervention, yet results in a recommendation for expensive, irreversible treatment.

The business model behind these tools adds a layer of complexity. Pearl, a leading manufacturer of dental AI, has highlighted its ability to increase treatment acceptance rates by 30%. When such metrics are used as primary sales arguments, it creates a potential conflict of interest.

The visual presentation of AI-generated, high-contrast, color-coded images on large screens can be highly persuasive, potentially anchoring the patient’s trust in the software over the clinician’s traditional diagnostic judgment.

The psychological phenomenon known as “machine heuristics” may exacerbate this issue. As noted by Jean-François Ouellet, a professor at HEC Montréal, there is a natural human tendency to perceive computer-generated data as inherently objective, precise, and impartial.

This bias can lead dentists to rely too heavily on algorithmic suggestions, even when those suggestions conflict with their own clinical experience or the patient’s specific health history. In a broader public health context, if AI adoption leads to a systemic increase in dental procedures, it could place significant, unplanned pressure on both private insurance premiums and public dental care funding.

The regulatory and ethical landscape remains in flux. While medical fiction—such as the recent portrayal in the Radio-Canada program Stat—has brought the risks of machine-led medical errors into the public consciousness, the practical, everyday reality is more nuanced. Dentists remain highly trained professionals bound by ethical codes and regulatory colleges, yet they are not immune to the influence of advanced technologies that promise both efficiency and increased revenue.

Contraindications and When to Seek a Second Opinion

Patients should exercise caution when presented with a significant, high-cost treatment plan that relies exclusively on AI-generated images without corroborating physical evidence or symptoms. If a dentist recommends multiple, aggressive procedures based solely on a software report, it is appropriate to request a manual review of the radiographs.

A second opinion is particularly warranted if the proposed treatment involves irreversible procedures, such as extractions or invasive surgical interventions, in the absence of pain or visible clinical decay.

Pearl, which was founded in 2019, has established itself as a pioneer in the field by securing the first-ever FDA clearance for AI capable of identifying diseases in dental x-rays. The organization has since expanded its regulatory footprint to 120 countries, cementing its position as a global leader in dental imaging analysis.

The company further distinguished itself by becoming the first and only dental AI provider to receive FDA clearance for both 2D and 3D radiologic image analysis. This expansion allows its platform to automate the identification of critical anatomical structures within 3D cone beam computed tomography scans, covering areas ranging from the maxillary sinus to the airway.

Ultimately, the role of AI should be to serve as a supportive diagnostic tool, not the final arbiter of medical necessity. As the technology continues to evolve, the burden remains on the patient to ask informed questions and on the dental profession to ensure that their reliance on digital assistants does not compromise the individualized care that characterizes high-quality medical practice.

Photo of author

Health Desk

Health Desk at Archyde.

Xiaomi Opens Pre-Sales for Smart Door Lock 5 Face Recognition in China

Cardinals Non-Tender Former First-Round Pick Nolan Gorman