The Use of Artificial Intelligence In Services for Children and Families: A Scoping Review of Applications and Ethical Considerations

Background: In recent years, artificial intelligence (AI) technologies have been increasingly applied across diverse fields within the social sciences. Service-oriented sectors, including child and family services, are adopting AI-driven approaches to improve service quality, efficiency, and accessibility. However, concerns remain regarding ethical implications, particularly in high-stakes decision-making contexts such as child welfare. This study presents a scoping review of the literature to examine how AI has been applied in services for children and families and to synthesize the associated ethical considerations.

Method: A scoping review was conducted of literature published between 2015 and 2025. Seven electronic databases were searched (PubMed, CINAHL, PsycINFO, ERIC, Social Services Abstracts, Sociological Abstracts, and Social Work Abstracts). A total of 3,359 records were identified through database searching and 7 additional studies through citation tracking. After removing duplicates, 2,686 studies were screened by title and abstract, and 111 full-text articles were assessed for eligibility. Ultimately, 22 studies met the inclusion criteria. A data template was used to extract information on data/sample, research methods, the use of AI, and ethical concerns.

Findings: Most included studies were conducted in the United States (n = 14), with others from European countries. The majority employed quantitative designs (77%) and large-scale datasets. Machine learning approaches dominated (82%), particularly predictive modeling, classification algorithms, and ensemble methods, often integrated with natural language processing for unstructured text data. Applications primarily focused on risk prediction and early identification of child and family vulnerability (80%), followed by decision support in child welfare systems, fairness enhancement, and automation of administrative processes. Most studies remained at development, testing, or validation stages, with limited real-world implementation. While AI generally demonstrated strong performance in prediction and classification tasks, outperforming traditional methods in many cases, persistent challenges were identified, including bias, limited interpretability, and difficulties in implementation within practice settings. Across studies, ethical concerns centered on algorithmic bias, structural inequalities embedded in data, lack of transparency and explainability, risks of misclassification, surveillance concerns, and accountability issues. A consistent theme was the importance of maintaining AI as a decision-support tool rather than a replacement for professional judgment.

Conclusion and Discussion: Overall, AI shows significant promise in enhancing predictive capacity and service efficiency in child and family services, yet its use remains ethically and practically constrained. The findings highlight the need for stronger fairness-aware model development, improved transparency and explainability mechanisms, and more robust evaluation in real-world practice settings. Future research should prioritize participatory and interdisciplinary approaches that involve practitioners, policymakers, and affected communities to ensure responsible AI integration. Clear ethical governance frameworks are also essential to balance innovation with the protection of vulnerable populations.