While a number of medical devices have been formally licensed for usage in children, there remains to date no guidance on the ethical use of pediatric data in artificial intelligence or machine learning (AI/ML) research.

To ensure fundamental ethical principles are prioritized through the ideation, development, iteration, deployment, and evaluation of AI/ML studies, researchers have highlighted the importance of formal ethics review and reporting procedures to improve safety and promote equity. These efforts must be inclusive of the pediatric community.

ACCEPT-AI is a framework of recommendations for safely including pediatric data in artificial intelligence and machine learning (AI/ML) research. It has been built on fundamental ethical principles of pediatric and AI research and incorporates age, consent, assent, communication, equity, protection of data, and technological considerations.

ACCEPT-AI has been designed to guide researchers, clinicians, regulators, and policymakers and can be utilized as an independent tool or adjunct to existing AI/ML guidelines.

Interests of children

While a number of medical devices have been formally licensed for usage in children, there remains to date no guidance on the ethical use of pediatric data in artificial intelligence or machine learning (AI/ML) research.

To ensure fundamental ethical principles are prioritized through the ideation, development, iteration, deployment, and evaluation of AI/ML studies, researchers have highlighted the importance of formal ethics review and reporting procedures to improve safety and promote equity. These efforts must be inclusive of the pediatric community.

Children and young people (CYP) under eighteen are underrepresented in research, with pediatric studies presenting age-specific challenges that span ethical, legislative, financial, and relational domains. Further, concerns about racial and gender disparities in pediatric research have been expressed, with calls to improve demographic reporting. These considerations must be accounted for in developing pediatric AI/ML technology.

What is ACCEPT-AI?

The generation of formalized guidelines, such as SPIRIT-AI and CONSORT-AI, have laid the foundation for the safe design, conduct, reporting, and early-stage evaluation of AI/ML studies.

The ACCEPT-AI framework highlights fundamental ethical considerations for pediatric data use in AI/ML research. At each stage of the AI life cycle, the framework promotes the evaluation and maximization of moral and ethical AI use by incorporating respect for persons, beneficence, non-maleficence, justice, transparency, and explainability in its recommendations.

Applications of ACCEPT-AI

There are numerous applications of ACCEPT AI. One instance is when a tertiary academic center enrolls pediatric patients in a study that involves the creation of an AI/ML algorithm for assessing vascular malformations of the face. Here, parental consent and subject assent are given importance. Because CYPs cannot legally consent for themselves, federal regulations include special protections for pediatric study subjects, including parental consent and assent of older pediatric subjects. The consent process must account for both chronological and developmental ages.

Yet another is an instance where importance is given to communication and equity. A recent qualitative exploration of twenty-one CYPs showed that they wished to contribute insights to the safe development of AI research. Age-appropriate communication is the cornerstone of pediatric practice. Therefore, all stakeholders must be provided with relevant information on the purpose and nature of proposed AI/ML studies and given examples of how their data may be utilized in the future.

Similarly, pediatric data must only be utilized when the data and technology address a clear need for the pediatric population. Researchers must be transparent about the needs and potential benefits for data use in their protocols and should clearly describe measures taken to minimize risk to pediatric subjects.

Furthermore, combining data across adults and children introduces age-related algorithmic bias and risks compromising a study's applicability, generalizability, and effectiveness, with potential impact on both populations. Clear documentation of the objective for which pediatric data will be collected and used in line with the ACCEPT-AI recommendations will help ensure key safety measures have been taken to avoid mixing of data unless there are clear indications to do so.

Conclusion

Pediatric populations face many challenges in healthcare and research settings. As the research community develops consensus guidelines for AI/ML algorithms and refines AI's ethical and moral use, specific protections for pediatric populations are essential.

 Legal protections and federal mandates regarding developing and deploying AI/ML algorithms for pediatric populations have also yet to be established.