Blockchain for Trustworthy Artificial Intelligence in Dentistry

Abstract

The race for developing and marketing the best inline artificial intelligence (AI) solutions is already in full swing in the dental industry. While regulators are trying to keep up with this fast-paced innovation, end users of these technologies must be on guard to navigate this new landscape safely. Trust is the foundation for this guardrail. Although regulatory approvals can provide some level of trust to an AI solution, users must be empowered with the knowledge of essential vocabulary and semantics to ask the right questions to assess the trustworthiness of the solution. This commentary elaborates on one technology proposed to build trustworthiness in AI solutions: blockchain. Further, we enlist a nonexhaustive list of questions for the users to ask when considering AI solutions in dentistry that may claim to use blockchain technology.

Knowledge Transfer Statement:

The topic discussed in this commentary could serve as an initial inquiry point that deeply probes into the trustworthiness of an AI solution that a user might consider applying in the field of dentistry.

The integration of artificial intelligence has seen a steep rise in several software solutions targeted toward dental clinic workflow and patient care. Almost all providers of these solutions are promising to revolutionize dental practice in diverse dimensions, including early and accurate diagnosis, patient education, and easy and precise documentation of claims. However, beyond the face value of these claims, how does a potential user trust these solutions regarding data privacy, security, ethical use of sensitive data, and truthfulness about performance gain? While regulatory approvals may ensure standardization in compliance with guidelines, some technologies can help with portraying the additional trustworthiness of an AI solution. Blockchain is one technology widely discussed as a solution for trustworthiness (Fritz 2022; Mokhamed et al. 2023). Simply put, blockchain technology avoids the need to trust the individuals or organizations that provide the AI solution; instead, the user only needs to trust the technology. Hence, it is sometimes referred to as a “trustless system.”

What Is Blockchain Technology?

Blockchain is a distributed digital ledger that records transactions across multiple computers, making the data immutable. The term blockchain comes from how the data is stored: in blocks linked together in a chain. Each block contains a collection of transactions, records, or data, and once a block is added to the chain, it becomes permanent. The immutability of blockchain refers to the fact that, once data is recorded, they cannot be altered without altering every other block in the chain.

How Does Blockchain Add Trustworthiness to AI Solutions?

Blockchain technology can address many of the trust-related concerns associated with AI solutions (Xiong et al. 2024; Zhang et al. 2024) in dentistry by offering the following advantages:

Data security and privacy: Blockchain’s decentralized nature ensures that data is stored securely across multiple nodes, reducing the risk of data breaches or unauthorized access. When sensitive patient data is stored or used in AI models, blockchain can help protect this data through encryption and decentralization, ensuring compliance with regulatory standards. In addition, the immutability of blockchain ensures that once patient records or AI model updates are recorded, they cannot be altered, providing confidence in data and model integrity.

Transparent and auditable AI model updates: AI models may undergo updates as new data become available or algorithm improvements are made. With blockchain, every update to an AI model can be recorded as a transaction (adding or removing information), creating a transparent and auditable log of all modifications. This means that dental professionals and regulatory bodies can easily track the history of changes to AI models, ensuring accountability on which version of the model was used to make a clinical decision.

Decentralized control: AI systems are often trained and tested on centralized databases controlled by a single entity (e.g., clinical data stored locally, in on-premises servers, or in the cloud). In contrast, blockchain operates on a decentralized network, meaning no single party has complete control over the data. This reduces the risk of data manipulation or censorship and allows for a more democratized approach to AI model development and validation. Thus, blockchain technology lends well to the federated learning paradigm in AI (Rischke et al. 2022).

Improved data traceability: Blockchain allows for complete data traceability, from the source of patient records to how this data is used in AI models. In the case of dental diagnostics, for instance, a dentist can trace the origin of imaging data and track its entire life cycle through the AI system. This transparency allows for greater trust in the system, as dental professionals can verify that the AI model uses accurate and high-quality data for training.

What Questions Should Potential Users Ask Providers That Claim to Use Blockchain-Based AI Solutions?

• 1. What specific components or features of your AI product leverage blockchain technology?

Blockchain can be applied in various ways (e.g., data storage, secure communication), and not all blockchain applications inherently guarantee security.

• 2. Can modifications to data be tracked transparently, and who has access to this information?

• 3. Who owns the data stored on the blockchain? How do you control access to the data for different parties, such as dentists, insurance providers, and patients?

Mismanaged access can lead to breaches of trust and legal violations. Only authorized individuals should be able to access or modify records, and patients’ data rights should be respected.

• 4. How scalable is the blockchain solution for handling large volumes of patient data? Will this affect the performance or speed of the AI solution?

The validation of new blocks often requires long computations and may not scale well with large amounts of data.

• 5. Is the blockchain solution compatible for data communication between other software I use?

• 6. What consensus mechanism (e.g., Proof of Work, Proof of Stake) does your blockchain use?

A consensus mechanism ensures the integrity of data on the blockchain. Knowing how decisions about data validation are made helps assess the security and efficiency of the system.

• 7. Are there any ongoing maintenance fees for using the blockchain network?

Blockchain can be expensive to implement and maintain. High operational costs could outweigh the benefits, especially for smaller practices.

• 8. How can clinical personnel, or patients, audit the blockchain to verify that all patient data have been securely handled?

Having the ability to audit the blockchain ensures transparency and trust in the system.

• 9. Has your blockchain solution undergone independent audits and certifications?

• 10. How does your blockchain implementation mitigate risks such as 51% attacks, data breaches, or intelligent contract vulnerabilities?

Even blockchain systems are not immune to attacks, such as 51% attacks or vulnerabilities in smart contracts. Asking probing questions on potential risks and vulnerabilities is essential for trust.

• 11. How does the blockchain handle data backup and recovery in case of system failures? What happens if the network fails or becomes inaccessible?

• 12. How does your solution manage patient consent? Is there a precise mechanism for patients to approve or revoke consent for their data to be used?

As provisioning dental care is rapidly becoming augmented with AI, these questions may provide a starting point for discussing the trustworthiness of these solutions.

Author Contributions

D. Cerda Mardini, S. Madathil, contributed to conception, design, and interpretation, drafted and critically revised the manuscript; M. Sharma, contributed to design, data interpretation and critically revised the manuscript. All authors gave final approval and agree to be accountable for all aspects of the work.