Empowering Conversations: The Potentials and Challenges of GenAI Chatbots in Sexuality Education

  • 14/02/2025
  • Category: Commentaries
No Comments

Author: Tjok Istri Sintawati
Editor: Ayom Mratita Purbandani

Comprehensive sexuality education (CSE) is a fundamental aspect of sexual and reproductive health and rights (SRHR) that must be fulfilled. Access to CSE equips individuals with better awareness of sexuality–cognitive, physical, and social aspects–while also preventing risks that might occur. Unfortunately, CSE implementation still encounters barriers due to the taboo, local cultures, and societal norms as happened in Indonesia.

With the growing development of technology, the utilization of generative artificial intelligence (GenAI)–which uses machine learning to process data and create new content–has become an alternative to overcome this problem through the application of chatbots. Currently, chatbots designed to provide sexuality education are starting to be developed e.g. SnehAI, PAT, AlexBot, and AdolescentBot. However, this alternative also comes with several challenges. This article discusses the potentials and challenges in developing GenAI-powered chatbots to facilitate CSE.

Lack of access to sexuality education and the presence of GenAI

CSE has proven to positively affect individual health in the long term1. Additionally, it plays an important role in the prevention of sexual and gender-based violence through deference to the formation of respectful social and sexual relationships and understanding of the rights of others2. A study by Maimmunah on 165 Indonesian adolescents (ages 12-19) found 73.8% of adolescents consider teaching their sexuality would make them better equipped to avoid high-risk sexual behaviour and sexual violence.3

In contrast to the urgency surrounding the issue, efforts to promote CSE encounter obstacles that impede its application. One of the main inhibiting factors is the continued prevalence of taboos around discussing SRHR. These taboos  are rooted in a historical, religious, and cultural context that gives broad implications to beliefs, behavior, health policies, and the educational programs4. As a result, both formal and informal settings, where the family transforms CSE in young people, become choked up.

As Daneback et al. explain, the internet and the media have then become a significant source of sex education for many young people5. Digital space helps overcome the discomfort of talking about the SRH of young people or teachers6. GenAI-powered chatbots, for instance, are believed to be a source of information about STIs, disease control, clinic search, and even sensitive questions relating to sexual activity7. GenAI-powered chatbots can generate answers about SRH and create outputs like text, image, audio, and even more. It obtains data by applying diverse methods both automatically and manually (e.g. web scraping and crawling, public data sets, crowdsourcing, synthetic data generation, customer data, up to user-generated content (UGC))8. Intended users from this GenAI-powered chatbots providing CSE are indeed young people themselves.9 10

Chatbots powered by GenAI for CSE have become numerous over time. One example is PAT chatbot, based in London and funded by the Public Health England HIV Innovation Fund, which can answer SRH-related questions and give sign-posts about HIV/STI prevention11. Other chatbots such as ANA Chatbot, AdolescentBot, and SnehAi are integrated in text messaging platforms such as Facebook Messenger and WhatsApp12. In addition, a well-known public chatbot like ChatGPT is also widely used by young people to access information about SRH. Scientists see this as a potential tool that is more accessible to teach sexual consent13.

Chatbot-powered sexual education challenges

Data bias and threats to inclusive content production

Applying GenAI in the chatbot as an alternative to many sex education currently developed is essentially a virtue. Research done by Greer et al. in 2024 shows that the use of ChatGPT to facilitate the exercise of sexuality education meets some criteria according to the US National Standards of Sex Education, such as accessibility, neutral responses and impartial information, and accuracy14. However, the relevance of information from the GenAI chatbots with its demographics–such as age, sex, race, education, family income–from young people who access it needs attention in further development.

A UNESCO study on large language models (LLM), which power GenAI, found that the LLM tends to produce gender bias, homophobic attitudes, and racial stereotyping15 These conditions can mismatch with needs and have implications that are relevant to the context experienced by vulnerable and marginalized communities. Potentially, preserving prejudice and stereotypes that can perpetuate the social exclusion of the group rather than achieve the initial intent of CSE.

Bias in the chatbot technology powered by GenAI can also result from the limited range of knowledge of the subject16. GenAI, like other AI technologies, gains knowledge from training in available data sets. Gaps in data availability may produce content that fails to meet the needs of different risks and populations17. For sexuality education to be effective and inclusive, the chatbots development should take place ethically.

Misinformation and disinformation

One significant risk faced by sexuality education through a chatbot powered by GenAI is related to misinformation and disinformation. Along with the opportunities to use the chatbot GenAI in sexuality education, the threat of manifest content might be misleading. There is at least some of the primary attention that puts GenAI at risk of backfiring to its users: an increase in quantity and quality of misinformation, increased personalization of misinformation, and involuntary but incorrect reproduction of information18.

For example, disinformation may involve the false information about forms of treatment or drugs which can be harmful if not thoroughly verified19. This issue might be caused by the unverified use of data sets to produce misleading content. As with testing done by Sharevski et al., the chatbot that trains unverified data can spread myths and misconceptions about contraception and abortion20.  Hence, the development of the chatbot for sexuality education requires direct monitoring from the SRH experts to ensure the quality of the data sets used. With that in mind, the accuracy of content produced by the chatbot powered by GenAI can be more optimal and minimize the misinformation and disinformation that may occur.

Conclusion 

GenAI has great potential as an alternative to facilitating CSE for young people, including those in Indonesia. However, unlocking its power requires addressing critical risks,  such as ethical and inclusive dimensions and threats of misinformation and disinformation. Equally important is safeguarding user data through security and transparent management of user data. By all this means, GenAI-powered chatbot could be a game-changer to empowering young people with the knowledge they deserve.

1.  Goldfarb, E.S. and Lieberman, L.D. (2021) ‘Three decades of research: The case for comprehensive sex education’, Journal of Adolescent Health, 68(1), pp. 13–27. doi:10.1016/j.jadohealth.2020.07.036. 

2. Comprehensive sexuality education (no date) World Health Organization. Available at: https://www.who.int/news-room/questions-and-answers/item/comprehensive-sexuality-education (Accessed: 26 January 2025). 

3.  Maimunah, S. (2019) ‘Importance of sex education from the adolescents’ perspective: A study in Indonesia’, Open Journal for Psychological Research, 3(1), pp. 23–30. doi:10.32591/coas.ojpr.0301.03023m. 

4.  Mohd. Tohit, N.F. and Haque, M. (2024) ‘Forbidden conversations: A comprehensive exploration of taboos in sexual and Reproductive Health’, Cureus [Preprint]. doi:10.7759/cureus.66723. 

5.  Daneback, K. et al. (2012) ‘The internet as a source of information about sexuality’, Sex Education, 12(5), pp. 583–598. doi:10.1080/14681811.2011.627739. 

6. Global Partnership Forum on Comprehensive Sexuality Education (2023). rep. Available at: https://healtheducationresources.unesco.org/library/documents/sexuality-education-digital-environment (Accessed: 26 January 2025). 

7.  Nadarzynski, T. et al. (2021) ‘Barriers and facilitators to engagement with Artificial Intelligence (ai)-based Chatbots for sexual and Reproductive Health Advice: A qualitative analysis’, Sexual Health, 18(5), pp. 385–393. doi:10.1071/sh21123. 

8.  Where do Generative AI models source their data & information? (2023) Smith.ai. Available at: https://smith.ai/blog/where-do-generative-ai-models-source-their-data-information (Accessed: 29 January 2025). 

9.  Revolutionizing sex education with AI-powered sexed chatbot (2025) AlexBot. Available at: https://alexbot.eu/ (Accessed: 29 January 2025). 

10. SnehAI. Available at: https://snehai.org/ (Accessed: 29 January 2025). 

11. Nadarzynski, T. et al. (2021) ‘Barriers and facilitators to engagement with Artificial Intelligence (ai)-based Chatbots for sexual and Reproductive Health Advice: A qualitative analysis’, Sexual Health, 18(5), pp. 385–393. doi:10.1071/sh21123.

12. Liew, T.W. et al. (2023) ‘Let’s talk about sex!: Ai and relational factors in the adoption of a chatbot conveying sexual and reproductive health information’, Computers in Human Behavior Reports, 11, p. 100323. doi:10.1016/j.chbr.2023.100323. 

13  Owsianik, J. (2024) What can chatgpt teach us about sexual consent?, LeShaw. Available at: https://leshaw.com/what-can-chatgpt-teach-us-about-sexual-consent/ (Accessed: 30 January 2025). 

14. Greer, K.M. et al. (2024) ‘Navigating the future of sexuality education in the USA: Applying technology mediation theory to AI-facilitated sexuality education’, Sex Education, pp. 1–15. doi:10.1080/14681811.2024.2401802. 

15.  UNESCO (2024) Challenging systematic prejudices: an investigation into bias against women and girls in large language models. rep. Available at: https://unesdoc.unesco.org/ark:/48223/pf0000388971 (Accessed: 26 January 2025). 

16.  Fetrati, H., Chan, G. and Orji, R. (2024) ‘Chatbots for Sexual Health Improvement: A Systematic Review’, International Journal of Human–Computer Interaction, pp. 1–23. doi:10.1080/10447318.2024.2376808. 

17.  WHO (2024) The role of artificial intelligence in sexual and reproductive health and rights. Available at: https://www.who.int/publications/i/item/9789240090705 (Accessed: 26 January 2025). 

18.  Simon, F.M., Altay, S. and Mercier, H. (2023) Misinformation reloaded? fears about the impact of generative AI on misinformation are overblown: HKS Misinformation Review, Misinformation Review. Available at: https://misinforeview.hks.harvard.edu/article/misinformation-reloaded-fears-about-the-impact-of-generative-ai-on-misinformation-are-overblown/ (Accessed: 27 January 2025). 

19.  Park, H.J. (2024) ‘The rise of Generative Artificial Intelligence and the threat of fake news and disinformation online: Perspectives from sexual medicine’, Investigative and Clinical Urology, 65(3), p. 199. doi:10.4111/icu.20240015. 

20.  Sharevski, F. et al. (2023) ‘Talking abortion (mis)information with CHATGPT on TikTok’, 2023 IEEE European Symposium on Security and Privacy Workshops (EuroS&PW), pp. 594–608. doi:10.1109/eurospw59978.2023.00071.  and expert digital workers into the workforce.