Evaluating the Role of Artificial Intelligence in Breast Self-screening (Preprint)

<sec> <title>BACKGROUND</title> With the increasing accessibility of artificial intelligence (AI), the general population is frequently engaging with conversational AI platforms to understand topics related to health and well-being. Breast cancer remains a significant global health issue, ranking as the second most common cancer in women and the leading cause of cancer-related mortality worldwide. While breast self-examination (BSE) is not a replacement for professional screening, it promotes body awareness and may support earlier recognition of changes. Evaluating AI’s role in patient education regarding such practices is therefore timely and important. </sec> <sec> <title>OBJECTIVE</title> This viewpoint aimed to explore the potential and limitations of AI-based conversational models in providing health advice on breast self-screening, using DeepSeek AI as an exemplar. </sec> <sec> <title>METHODS</title> A qualitative review was conducted of a dialogue session with DeepSeek AI, where six commonly asked questions regarding breast self-examination were posed. Responses were assessed in comparison with guidance from established health organisations such as the World Health Organization (WHO) and Breastcancer.org. The evaluation focused on accuracy, comprehensibility, and alignment with evidence-based practices. </sec> <sec> <title>RESULTS</title> DeepSeek AI generated comprehensive and accessible responses, employing language suitable for the general population and maintaining consistency with professional recommendations. The model explained BSE techniques clearly, acknowledged alternative approaches, and consistently emphasised that self-examination is complementary rather than a substitute for medical screening. However, notable limitations included insufficient emphasis on the limited mortality benefits of BSE, lack of references to individual variability (e.g., dense breast tissue), and absence of links to credible video resources, which may enhance understanding for visual learners. </sec> <sec> <title>CONCLUSIONS</title> AI-driven dialogues demonstrate promise in reinforcing public health practices such as breast self-examination, offering accurate and comprehensible information. Nevertheless, oversimplification, inadequate contextualisation, and lack of integration with multimedia resources remain challenges. For responsible use in health education, AI responses should be supplemented by physician oversight, structured public health campaigns, and clear signposting to diagnostic services. Strengthening AI outputs with evidence-based statistics and visual learning resources can optimise their role as supportive tools in preventive health. </sec> <sec> <title>CLINICALTRIAL</title> N/A </sec>