Pushy AI chatbots risk putting patients off screening appointments
The rise of AI chatbots in healthcare is an exciting prospect, but new research from the University of Surrey highlights a potential pitfall: over-messaging and pushy reminders can turn patients off. The study, published in Lingua, reveals that how an AI chatbot communicates is just as important as what it does. Patients who responded positively to the AI described its tone as friendly and kind, appreciating its integration into their routines. However, the research also uncovered clear friction points, particularly around follow-up messages and imperative phrasing.
One of the key findings was the impact of anthropomorphism. Patients who felt the AI was too human-like and aggressive, especially with imperative phrasing like "Let's book you in," found it off-putting. This was particularly challenging for those managing mental health issues, neurodivergent conditions, or demanding caring responsibilities, as the pressure to respond quickly felt unfair.
Ethical concerns were another significant issue. Patients were wary of data security, impersonation, and the blurred lines between human and AI interactions. The chatbot's statement that users could "chat to me as if I am a real person" was seen as suspicious rather than reassuring, undermining trust.
Dr. Doris Dippold, lead author of the study, emphasizes the importance of transparency and respect in chatbot design. She states, "Human-like features can build rapport, but when they clash with patients' expectations for transparency in a healthcare setting, they undermine the trust the chatbot aims to establish."
The study's findings carry weight, especially considering the declining cervical screening uptake in the UK. The research underscores the need for healthcare chatbots to be designed with principles such as helping people achieve their goals, giving them control over decisions, responding appropriately to needs, ensuring fairness, and being transparent about their technology.
In a diverse and socioeconomically deprived community, where equitable communication is crucial, the study highlights the importance of building rapport through respectful, transparent, and supportive communication. Dr. Dippold concludes, "Feeling seen, appreciated, and emotionally supported is not a luxury feature in health AI; it is a condition of access."
This research serves as a reminder that while AI chatbots have the potential to revolutionize healthcare, their design and implementation must be carefully considered to ensure patient engagement and trust.
