Dies ist eine Übersichtsseite mit Metadaten zu dieser wissenschaftlichen Arbeit. Der vollständige Artikel ist beim Verlag verfügbar.
ChatGPT as a tool to improve access to knowledge on sexually transmitted infections
8
Zitationen
4
Autoren
2024
Jahr
Abstract
OBJECTIVES: Specific to sexual health, individuals in need of information may be adolescents who have limited ability to formally access healthcare. These digital natives may turn to ChatGPT to address their concerns on sexually transmitted infections (STI). We sought to evaluate the veracity of ChatGPT's responses to commonly asked questions on STIs. METHODS: We instructed ChatGPT (GPT 3.5) to answer STI questions from three domains, namely, (1) general risk factors for STIs, (2) access to care and diagnosis of STIs and (3) management of STIs and postexposure prophylaxis. The responses were recorded and checked against the US Centers for Disease Control and Prevention STI Treatment Guidelines 2021. RESULTS: Overall, the responses were concise and accurate. In terms of prevention, ChatGPT could also recommend measures like safe sex practices and human papillomavirus vaccination. However, it failed to recommend HIV pre-exposure prophylaxis. When an individual expressed a symptom that could potentially represent STI (eg, dyspareunia) ChatGPT appropriately provided reassurance that other possibilities exist, but advocated for testing. In terms of treatment, ChatGPT consistently communicated the importance of partner testing and follow-up testing, but at times, failed to highlight the importance of testing for other STIs. Overall, the advice given was not tailored to the specific individual's circumstances. CONCLUSIONS: ChatGPT can provide helpful information regarding STIs, but the advice lacks specificity and requires a human physician to fine-tune. Its ubiquity may make it a useful adjunct to sexual health clinics, to improve knowledge and access to care.
Ähnliche Arbeiten
Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI
2019 · 8.635 Zit.
Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead
2019 · 8.543 Zit.
High-performance medicine: the convergence of human and artificial intelligence
2018 · 8.051 Zit.
BioBERT: a pre-trained biomedical language representation model for biomedical text mining
2019 · 6.844 Zit.
Proceedings of the 19th International Joint Conference on Artificial Intelligence
2005 · 5.781 Zit.