A study published in the journal Lingua analyzed patient interactions with a generative AI receptionist used to invite patients to book cervical screening appointments. The research, conducted by the University of Surrey, found that specific elements of the chatbot’s communication style affected patient trust and engagement, with both positive and negative responses reported.
Research Methodology
The study examined the performance of a chatbot named 'Asa,' designed to book cervical screening appointments via WhatsApp. The research was based on interviews at a single GP surgery in North London and a survey of 300 people who were eligible for cervical screening through the NHS. The GP surgery in Islington was described as serving a highly diverse, socioeconomically deprived community.
Key Findings
Positive Responses
Patients described the chatbot’s tone as "friendly," "kind," and "not forceful." Some patients reported that it was easier to disclose sensitive information—such as cancelling an appointment due to menstruation—to a named, female-presenting AI compared to speaking with a male receptionist. The integration of the chatbot into existing routines was also cited positively.
Negative Responses
Follow-up messages sent within 24 hours were perceived as intrusive. Imperative phrasing, such as "Let's book you in," was described as aggressive. Patients with mental health challenges, neurodivergent conditions, or caring responsibilities reported feeling pressured by expectations of a quick response. The chatbot's directive "chat to me as if I am a real person" was interpreted as suspicious rather than reassuring.
Ethical Concerns
The most consistently reported negative feedback related to ethical issues, including data security, potential impersonation, and blurred boundaries between human and AI interaction.
Study Recommendations
The researchers suggested that healthcare chatbot design should adhere to several principles:
- Helping users achieve their goals
- Giving them control over decisions
- Responding appropriately to their needs
- Treating them with respect
- Ensuring fairness
- Being transparent about how the technology works
Statements from Lead Author
Dr. Doris Dippold, lead author and Associate Professor at the University of Surrey, stated:
"Our analysis shows that anthropomorphism is not universally positive. Human-like features can build rapport, but when they clash with patients' expectations for transparency in a healthcare setting, they undermine exactly the trust the chatbot is trying to build."
Dr. Dippold also said:
"Feeling seen, appreciated and emotionally supported is not a luxury feature in health AI - it is a condition of access. If patients disengage because a chatbot feels pushy or untrustworthy, the health service loses them entirely."
Background on Cervical Screening
The study was conducted against the backdrop of declining cervical screening uptake, which fell 5.3% in the UK in 2023-24. Officials have noted that ethnic minority groups are underrepresented in screening participation.