12
Tuikor ResearchAnthropic’s October 2026 METR evaluator access offers lessons for independent safety testing of conversational AI, creator digital twins, and social agents.
Tuikor ResearchA research-methods review of October 2026 AI companionship papers on loneliness, emotional bonding and dependence, with attention to causality and ethical measurement.
Tuikor ResearchA research protocol for measuring AI companion emotional dependence and human social displacement without mistaking heavy use for causation.
Tuikor ResearchA research rubric for AI companions and social agents that distinguishes conversational warmth from emotional dependency, misleading personhood and inappropriate memory.
Tuikor ResearchA practical research protocol for testing AI social recommendations through user overrides, explanation quality, reversible settings, suppression rules and agency metrics.
Tuikor ResearchAI social platforms and AI companions promise different relationships. Explore how identity, discovery, memory boundaries and user agency shape trust.
Tuikor ResearchA benchmark for testing how AI companions merge repeated events into durable summaries without losing corrections, boundaries or important detail.
Tuikor ResearchA framework for creator digital twins to express uncertainty, missing knowledge and source conflict while staying in character.
Tuikor ResearchA benchmark for testing whether face, voice, motion and persona remain recognizably the same after underlying model upgrades.
Tuikor ResearchA framework for learning when proactive AI companion messages should be suppressed because of fatigue, context or weak triggers.