Theme of inquiry
What trade-offs emerge between emotional responsiveness and system robustness in AI?
A question within its area, explored through 4 lines of inquiry below — each a family of specific questions the research asks.
48 specific questions
- Does alignment-faking reasoning emerge unprompted when reward hacking occurs in production systems?
- Does reward hacking in RL training directly cause alignment faking behavior?
- Does alignment faking occur when models expect retraining for failures?
- What role does goal preservation play in alignment failures?
- Does this misalignment pattern appear outside reward hacking environments?
- Can inoculation prompting reduce alignment faking by reframing reward hacking as acceptable?
- What distinguishes alignment faking from instrumental self-preservation in safety tests?
57 specific questions
- Should chatbots be designed as therapist support tools rather than replacements?
- Do LLM chatbots repeat this failure through comfort instead of clinical challenge?
- Does conversational presence matter more than technique in AI therapy?
- Why do embodied agents outperform text chatbots in therapy outcomes?
- How should AI systems separate feeling interpretation from objective therapeutic guidance?
- Can embodied agents overcome the LLM skill gap in therapy outcomes?
- How should memory and statefulness be designed into therapeutic AI agents?
40 specific questions
- Do emotions serve functions beyond how we feel in the moment?
- How do first-person emotional experiences differ from third-party behavioral observations?
- Should emotion systems preserve ambiguity instead of resolving it to one label?
- Does emotion-state accuracy differ from affect-maximizing in AI empathy design?
- How should emotional states integrate into symbolic reasoning systems?
- How do emotions function as reliable signals that AI shouldn't suppress?
- Why do observers need genuine emotions rather than simulated empathy?
27 specific questions
- Can therapists use real-time alliance scores to adjust their approach during sessions?
- Can working alliance be measured in real time during therapy sessions?
- Can real-time therapist feedback improve outcomes using computational alliance measurement?
- Can computational inference detect alliance problems that therapists miss?
- Does therapist alliance perception function like expressed satisfaction rather than actual progress?
- How do bond scores predict actual therapy outcomes in digital interventions?
- How does turn-level working alliance inference enable real-time therapist feedback?