Can Medical Chatbots Trigger Disinhibition and Encourage Health Information Disclosure?

Area of Science:

• Human-Computer Interaction
• Health Communication
• Artificial Intelligence in Healthcare

Background:

• Medical chatbots are increasingly used to improve patient communication and reduce stigma.
• The assumption that chatbots foster disclosure of sensitive health information lacks consistent empirical support.
• Online disinhibition theory is explored to understand interactions with AI agents.

Purpose of the Study:

• To introduce and examine the concept of machine-mediated disinhibition (MMD).
• To compare disclosure levels in chatbot, human-through-computer, and face-to-face interactions.
• To investigate whether chatbot consultations increase patient willingness to disclose sensitive health data.

Main Methods:

• A scenario-based, between-subjects experimental design was employed.
• Three interaction modes were compared: face-to-face, human-through-computer, and chatbot.
• The study involved 373 participants.

Main Results:

• No evidence of increased disinhibition was found in the chatbot interaction condition.
• Participants demonstrated significantly lower willingness to disclose sensitive health information to chatbots compared to humans.
• Trust-related concerns appeared to override potential disinhibition effects.

Conclusions:

• In high-stakes healthcare settings, trust issues with AI may lead to information avoidance rather than openness.
• The study challenges the assumption that AI agents inherently promote patient disclosure.
• Further research on trust in AI-mediated medical communication is critical.