Frequently Asked Questions - Artificial Intelligence in Dentistry

Yes, the AI in Dentistry Guidance applies to all types of AI tools; however, dentists may need to exercise greater caution and oversight depending on the specific tool. AI tools pose a higher risk if they involve personal health information (including tools that are integrated with patient records for administrative or operational tasks) and/or impact clinical care (e.g., tools that influence clinical decision-making by helping formulate a diagnosis or treatment plan).

Steps dentists can take to ensure that an AI tool is safe and secure include: 

• Review the company’s policies and ask about their compliance with relevant privacy law (e.g., the Personal Health Information Protection Act, 2004, or Personal Information Protection and Electronic Documents Act), including how a tool collects, uses, and discloses data, retention practices, and security measures;
  • The RCDSO’s Electronic Records Management Guidelines caution dentists around the use of cloud service providers and the associated privacy and security risks. When selecting AI tools, the location of where data is stored can impact how it may be used (e.g., jurisdictions outside of Canada may have privacy laws that may allow broader uses of de-identified health information without consent.)
• Conduct a risk-based assessment and privacy impact assessment to understand potential risks and consequences;
• Consider cybersecurity insurance, especially if AI tools are integrated with the records management system;
• Follow relevant guidance from the Office of the Privacy Commissioner of Canada or Information and Privacy Commissioner of Ontario (e.g., on the use of generative AI or privacy management for small health care organizations); and
• Seek legal advice to understand how to protect personal health information of patients in accordance with the law.

Health Canada approval indicates that AI manufacturers of an AI-enabled medical device have met standards to demonstrate the safety and efficacy of their device through the product lifecycle, including those related to design, risk management, data selection and training, testing and evaluation, clinical validation, and transparency. Manufacturers are also asked to demonstrate device safety and effectiveness in diverse patient populations. For more information, see Health Canada’s Pre-market guidance for machine learning-enabled medical devices.

While many AI-enabled medical devices have been approved, not all AI tools qualify as medical devices. For instance, administrative tools and AI scribes may not be regulated as medical devices, but dentists may still choose to use them in their practice. When considering these tools, dentists can review the assessment criteria outlined in the AI in Dentistry Guidance to decide whether the tools are appropriate to use.

No, dentists are not required to use AI in their practice. AI is not required for dentists to provide competent and quality care to patients. Dentists may use AI tools as part of their practice if doing so aligns with or supports the patient’s best interests, such as by improving the delivery and quality of patient care.

AI scribes can be a tool to help reduce administrative burden and improve documentation. Some key considerations when using an AI scribe include:

• Obtain patient consent: Before using AI scribes, it is important to obtain patient consent since a patient’s personal health information is involved. Use clear and plain language with patients to explain the purpose and potential uses of any recordings. OntarioMD’s Patient Consent Toolkit provides examples of how to obtain patient consent for the use of AI scribes.
• Checking for accuracy: Consider whether the tool is trained to recognize relevant terminology and process diverse accents that reflect the patient population. Dentists are ultimately responsible for their documentation, so it is critical to always review AI-generated notes to ensure that they are comprehensive and accurately reflect patient encounters.
• Data processing, storage, and retention: To evaluate the privacy and security of the tool, find out whether recordings used to generate the transcription are created and retained by the tool, as well as where data are stored. Refer to RCDSO’s Electronic Records Management Guidelines and Dental Recordkeeping Guidelines for more information on security, retention, and disposal of clinical records.

OntarioMD has developed legal and privacy resources for healthcare providers using AI scribes. Although developed primarily with physicians in mind, dentists may also wish to review the Ontario AI Scribe Program for additional selection criteria.

As stated the AI in Dentistry Guidance, dentists are responsible for their clinical care and decision-making when using AI tools. While AI tools can help with many administrative, operational, and clinical tasks, they can still produce inaccurate, incomplete, or misleading results. Dentists are expected to exercise their professional judgement and override AI-generated outputs as necessary and appropriate. It is not appropriate for dentists to rely solely on AI tools without actively overseeing and critically reviewing AI-generated outputs to ensure they are accurate, complete, and reflect the patient’s individual circumstances.

As stated in the AI in Dentistry Guidance, dentists using AI in their practices remain responsible and accountable for patient care, clinical decision-making, and documentation. While AI can create efficiencies and support quality care, AI-generated outputs can contain errors and inaccuracies. It is important for dentists to mitigate these risks by providing appropriate oversight and critically reviewing AI-generated outputs.

It is not appropriate for a dentist to rely solely on AI to make clinical decisions. Dentists may also be held responsible for acts of professional misconduct under the Professional Misconduct regulation which can occur while using AI, for example, if they recommend or provide an unnecessary dental service.

Determining legal liability in civil law (i.e., when being sued) can be complex, given the potential role of AI manufacturers, developers, and institutions in the implementation of AI. Determining legal liability is beyond the scope of the AI in Dentistry Guidance and the role of the RCDSO. If dentists have specific questions relating to civil liability, they may wish to seek a lawyer with relevant expertise or contact the Professional Liability Program.

Dentists using AI tools in their practice are encouraged to provide feedback based on their experiences to AI vendors to help improve the tools. In particular, significant gaps or errors which may impact clinical care and decision-making are important to report.

For machine learning-enabled medical devices, Health Canada guidance on reporting a medical device problem is available here.

Dentists must also report incidents involving privacy breaches to the Information and Privacy Commissioner of Ontario (IPC). For more information, see the IPC’s website here.

To provide appropriate oversight when using AI, dentists require the knowledge and competence to identify when an AI-generated output may be wrong, inaccurate, or incomplete. The specific education and training dentists need will usually depend on the AI tool being implemented. Dentists can seek out tool-specific training from the AI vendor of the tool they intend to implement in their practice. Dentists are also encouraged to pursue relevant learning opportunities, including reading peer-reviewed evidence-based literature on the use of AI in dentistry, and completing continuing education courses approved by RCDSO and trusted sources, such as approved sponsors, academic institutions, and established online learning platforms.

The development and use of AI tools can have a significant environmental impact. Dentists interested in using AI sustainably can ask AI companies about the energy resources used to train and operate their tools, as well as the energy and water intensity of their technologies, to make more informed choices. AI has a potential role in finding solutions to environmental problems, such as by creating efficiencies and assisting with virtual care, which can reduce waste associated with travelling or using dental supplies and equipment for in-person visits.

As the use of AI in dentistry is new and evolving, a proactive approach towards disclosure can help foster patient trust and confidence in dentists and their use of AI tools in their care. Disclosure supports patient autonomy and informed decision-making. Some AI tools can influence a dentist’s decision-making, for example, by suggesting diagnoses and treatment recommendations, and therefore impact patient care and outcomes more directly than traditional tools. This potential impact on a patient’s care makes it more important to inform patients about the use of AI tools so that they can make informed decisions about their care. Patient consent is also required where personal health information is involved and is used by an AI tool, and when shared with third parties.

The AI in Dentistry Guidance recommends that dentists disclose to patients when they are using AI for clinical care. How dentists decide to inform patients is up to their discretion(e.g., consent form, signage, automated message, discussion with the patient) and can vary based on the nature and use of the tool. OntarioMD’s Patient Consent Toolkit provides examples of how to obtain patient consent in the context of using AI scribes). Patients may have different levels of interest in and understanding around AI tools. Dentists can consider sharing what tool was used, how it works, its risks and limitations, what safeguards are in place, and how it helped inform clinical decisions.

The RCDSO’s Dental Recordkeeping Guidelines apply to documentation created with the support of AI. Consistent with general guidance for effective recordkeeping, it is important to record details of any relevant discussions with patients around their care and the use of AI, including questions or concerns expressed by the patient. In line with a transparent approach to the use of AI, dentists can also include a notation to indicate that AI has been used to generate documentation or to support clinical decisions.

When using an AI tool that makes clinical recommendations (e.g., a diagnostic tool), dentists may choose to document their rationale for agreeing or disagreeing with the AI-generated recommendations based on their own assessment of the patient, knowledge, and professional judgment.

There may be patients who have concerns, are unfamiliar with, or feel uncomfortable with the use of AI. Dentists build trust and respect and support patient autonomy by taking reasonable steps to accommodate a patient’s wishes around the use of AI, to the extent this is possible and does not interfere with the provision of care. Where dentists require patient consent for the use and disclosure of patient’s personal health information in an AI tool, dentists must not use AI tools without such consent.

Exploring a patient’s concerns can clarify why they may not want AI to be used. Dentists may wish to provide patient education and reassurance, including around the dentist’s ultimate responsibility for decision-making, and the safety and security of the patient’s personal information.  Consistent with general guidance for effective recordkeeping, it is important for dentists to record details of relevant discussions with patients around their care, including questions or concerns expressed by the patient around the use of AI.