The integration of AI chatbots in healthcare is an exciting prospect, but a recent study from the University of Surrey reveals a potential pitfall: patients' perceptions of these chatbots can significantly impact their willingness to engage with important medical services. The research, focusing on cervical screening appointments, highlights the importance of a delicate balance between friendliness and transparency in chatbot design.
The Power of Friendliness
The study found that patients responded positively to a friendly and choice-oriented tone from the AI chatbot, Asa. Patients appreciated the system's ability to integrate into their routines and the ease of disclosing sensitive information, such as canceling appointments due to menstruation, with a female-presenting AI.
This human-like approach can build rapport and trust, which is crucial in healthcare. However, the research also uncovered some potential pitfalls.
Over-Messaging and Pushiness
Many patients found the follow-up messages sent within 24 hours to be intrusive and the imperative phrasing like "Let's book you in" to be aggressive. For individuals managing mental health challenges or neurodivergent conditions, the pressure to respond quickly felt unfair.
This over-messaging and pushiness can create a sense of urgency and anxiety, potentially deterring patients from engaging with the chatbot.
Ethical Concerns and Trust
The most significant issue, however, was the ethical concerns surrounding data security, impersonation, and the blurring of human-AI boundaries. Patients were suspicious of the chatbot's statement that they could "chat to me as if I am a real person," which undermined the trust the chatbot aimed to build.
Dr. Doris Dippold, the lead author of the study, emphasizes that anthropomorphism is not universally positive. When human-like features clash with patients' expectations for transparency, they can erode trust rather than build it.
Designing Trustworthy Chatbots
The study suggests that healthcare chatbots should be designed with the following principles in mind:
- Helping patients achieve their goals
- Giving patients control over decisions
- Responding appropriately to individual needs
- Treating patients with respect
- Ensuring fairness
- Being transparent about the technology's workings
These principles are particularly important in a diverse and socioeconomically deprived community, where equitable and accessible communication is a priority.
Conclusion: Building Trust, Not Just Appointments
The study highlights the need for healthcare chatbots to build rapport through respectful, transparent, and supportive communication. Patients must feel seen, appreciated, and emotionally supported, as this is a condition of access, not a luxury feature.
If patients disengage due to a chatbot's pushiness or untrustworthiness, the healthcare system loses a valuable connection with its patients. This research serves as a reminder that the design of healthcare chatbots should prioritize trust and transparency to ensure effective patient engagement and care.
