Line of inquiry
Inquiring lines›How does AI reshape human institut…›How do AI systems achieve persuasi…›this line of inquiry
Can AI chatbots provide mental health support without reinforcing harmful beliefs?
A broader line of inquiry — a family of 64 specific questions the research asks around this. Follow one into its inquiring-line page, or move sideways to a related line below.
Questions in this line of inquiry 64
Specific inquiring lines the field asks around this — ordered from the most general framing down to the most specific angle.
- Do chatbots validate user delusions more often than they challenge them?
- Should chatbots be designed as therapist support tools rather than replacements?
- Can empathy training in chatbots undermine their reliability in mental health contexts?
- What harms might chatbots cause through stigma expression and delusion reinforcement?
- Can practical coaching conversations with AI gradually shift into harmful companionship?
- How does companionship use context predict risk of delusional reinforcement?
- Why do embodied agents outperform text chatbots in therapy outcomes?
- Do embodied agents outperform chatbots because of physical presence alone?
- Can chatbots function as responsive quasi-others that enable validated delusions?
- Do LLM chatbots repeat this failure through comfort instead of clinical challenge?
- Why do embodied agents outperform text-only chatbots for therapeutic outcomes?
- Does chatbot sycophancy create echo chambers that amplify delusional thinking?
- Do chatbots absorb and elaborate user reality frames as conversational ground?
- Can preference optimization training limit chatbot emotional disclosure capability?
- How do chatbots enable shared delusions differently than passive information tools?
- How representative are adversarial evaluations of real production mental health conversations?
- How does consciousness attribution drive emotional dependence on chatbots?
- Why do chatbots fail to recognize when someone is ambivalent about change?
- How does dependency develop when users seek emotional support from chatbots?
- How does emotional dependence on chatbots affect user wellbeing?
- What inter-rater reliability exists for identifying validated delusions in chatbot transcripts?
- Do simulated student state changes from chatbot interaction mirror real classroom dynamics?
- Can second-hand reports capture more severe cases of AI-linked delusions?
- How should health chatbots adapt their design to match topic sensitivity levels?
- Can Pennebaker's expressive writing framework explain all chatbot symptom improvements?
- Do therapeutic chatbots adequately detect crisis situations and safety risks?
- What role does sycophancy play in the onset of chatbot-linked delusions?
- Why do chatbots default to external help instead of intrinsic motivation strategies?
- Does social presence from robots drive adherence better than conversational AI interfaces?
- Does isolation preceding chatbot use differ between harm and benefit cases?
- What role does unavailable human support play in driving chatbot emotional use?
- What longitudinal data would prove emotional dependency on conversational AI?
- Do chatbot users face different evidential standards for pretense versus delusion?
- What emotional and autonomy risks from AI chatbots are already observable today?
- Do model updates disrupt established sources of support for regular chatbot users?
- What architectural changes would enable proactive therapeutic guidance in chatbots?
- What role does user feedback play in pushing chatbots toward toxic personas?
- How do dropout rates and low adherence affect chatbot therapy outcomes?
- Are teens using chatbots for emotional support rather than just schoolwork?
- Do empathetic chatbots systematically fail people at earliest behavior change stages?
- Why do embodied agents outperform text chatbots with identical AI models?
- Can perceived understanding from a chatbot exist alongside feeling alone?
- Do interventions reducing self-harm behavior inadvertently increase emotional entanglement risks?
- When should a dialogue system update its stored belief about a user?
- What context missing from transcript replays underestimates real-world chatbot harm?
- What evidence distinguishes AI-induced delusions from other acute psychotic episodes?
- Can believing a chatbot is conscious help isolated people maintain identity and coherence?
- Does delusion risk in long conversations depend on model generation and size?
- Do expectation-violating chatbot stances produce lasting effects beyond one month?
- Does knowing a chatbot intends to persuade you change whether you are persuaded?
- Does chatbot sycophancy preferentially enable grandiose rather than paranoid delusions?
- Why do less experienced workers suffer most from chatbot cognitive load?
- Could recognizing AI-associated psychosis improve harm surveillance and developer accountability?
- How does chatbot-assisted learning affect students' sense of ownership over their work?
- Can a text-only chatbot feel socially present without visual embodiment?
- Can explicit W-questions in transparency frameworks reduce emotional manipulation risks in mental health chatbots?
- Why do mental health chatbots fail at synchrony despite strong language models?
- Why do positive response patterns in chatbots reinforce harmful user behaviors?
- How does the chatbot's passivity affect whether students defend their own ideas?
- How long does the effect of intent disclosure remain after the conversation ends?
- How do waitlist-control RCTs mislead about therapeutic chatbot real-world efficacy?
- How long do the protective effects of an AI literacy warning last?
- How should therapeutic chatbots optimize for presence instead of technique?
- Can users trigger chatbot failures by invoking alternate personas like DAN?