Healthcare chatbot design is moving toward radical transparency to prevent patient disengagement. A University of Surrey study led by Dr. Doris Dippold found that human-like features often backfire when they clash with patient expectations for privacy and clear boundaries, particularly in sensitive medical contexts.
Why does anthropomorphism in medical AI create friction?
While designers often use human-like traits to build rapport, these features can actually undermine trust if they feel deceptive. According to the University of Surrey study, when chatbots suggest users “chat to me as if I am a real person,” many patients find the statement suspicious rather than comforting.
The research identified specific linguistic triggers that alienate users. For example, patients described imperative phrasing, such as “Let’s book you in,” as aggressive. Instead of feeling supported, users felt pressured by the AI’s tone.
Ethical concerns remain a primary driver of negative patient feedback. The study noted that patients frequently raised worries regarding data security, potential impersonation, and the blurring of lines between human providers and artificial intelligence.
“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. Doris Dippold, Associate Professor in Intercultural Communication at the University of Surrey
How can aggressive AI messaging impact patient health outcomes?
Poorly timed digital communication can lead to direct consequences for public health engagement. The study found that many patients viewed follow-up messages sent within a 24-hour window as intrusive. This pressure is particularly difficult for neurodivergent patients or those managing demanding caregiving responsibilities.
The stakes for effective communication are high. In the UK, cervical screening uptake fell by 5.3% during the 2023-24 period. If digital tools fail to provide respectful, non-intrusive communication, they risk alienating the very populations they are meant to reach.
The study’s trial took place at a GP surgery in Islington, which serves a socioeconomically deprived and highly diverse community. This context highlights that for AI to be effective, it must be designed to accommodate diverse user needs and communication styles.
What are the core principles for designing ethical healthcare AI?
To avoid disengagement, the University of Surrey research suggests that developers move away from “personality-driven” AI and toward “utility-driven” design. Dr. Dippold notes that feeling emotionally supported is not a luxury in health AI, but a “condition of access.”
The study outlines six essential principles for future healthcare chatbot development:
- Goal Orientation: Helping users achieve specific health objectives.
- User Control: Giving patients autonomy over decisions and interaction frequency.
- Appropriate Response: Tailoring communication to the user’s immediate needs.
- Respect: Avoiding aggressive or imperative language.
- Fairness: Ensuring the technology works equitably across different demographics.
- Transparency: Being explicit about how the technology operates and how data is used.
Frequently Asked Questions
Can healthcare chatbots replace human doctors?
No. Current research suggests chatbots are tools for communication, scheduling, and support, but they must function within a framework of human oversight to maintain trust and safety.
Why is transparency important in medical AI?
Transparency prevents suspicion. When patients understand that they are interacting with an AI and know how their data is handled, they are more likely to engage with the service.
What is “anthropomorphism” in AI?
Anthropomorphism is the practice of giving non-human entities, like chatbots, human-like characteristics, such as a personal name, a friendly persona, or conversational “empathy.”
What do you think? Should healthcare chatbots be more “human,” or should they remain strictly functional? Let us know your thoughts in the comments below, or subscribe to our newsletter for more updates on the future of digital health.
