Keywords: primary health care, artificial intelligence, health equity, health services

Background:

Integrated primary healthcare (PHC) ensures that all individuals have access to health services. Artificial Intelligence (AI) has significantly transformed PHC by enhancing the quality, efficiency, and reach of care. However, concerns have emerged regarding AI’s potential to either reduce or exacerbate health inequities, one of the most persistent challenges in healthcare systems globally. Understanding AI’s role is essential for equipping the next generation of family physicians with the knowledge and insights needed to lead equitable and innovative care.

Research questions:

The primary objectives of the research revolve around two main questions: (a) how AI affects health equity in the PHC setting, and (b) what the contribution of AI in PHC to health inequalities is.

Method:

A scoping review was conducted with literature research across PubMed, Scopus, IEEE Xplore databases, and grey literature sources such as JSTOR and Google Scholar, covering the period from 2000 to 2025. Article selection adhered to the PRISMA-ScR guidelines, and thematic analysis was used to synthesise findings.

Results:

Out of 1,211 identified publications, 25 met the inclusion criteria. The results were categorized into eleven thematic domains, reflecting both positive and negative impacts of AI on health equity: (1) improving access to healthcare and addressing the digital divide, (2) enhancing early disease detection in underserved populations, (3) reducing disparities in clinical decision-making, (4) agency for self-care, (5) algorithmic bias, (6) ethical concerns, (7) patient trust, (8) dehumanisation and biomedicalization, (9) patient-doctor relationship, (10) participatory approaches and community involvement, and (11) provider acceptance, opportunity loss, and equity.

Conclusions:

This review summarises the extent to which implementation of AI in PHC promotes health equity or mitigates health inequalities and highlights the urgent need for further research to ensure its equitable implementation in healthcare systems and better prepare and empower future family physicians to navigate and lead in a rapidly transforming healthcare environment.

Points for discussion:

AI systems are only as fair as the data and assumptions behind them. Biases in training data or model design can lead to unequal treatment, especially for marginalized populations. What safeguards should be in place to prevent such biases from affecting care decisions?

How can family physicians critically evaluate and challenge algorithmic outputs, especially when they conflict with clinical judgment or patient context?

How can we train and empower future physicians to critically engage with AI tools while preserving the human touch in care?