A poll on Hacker News has gathered community perspectives on how to determine if an AI is sentient. One participant proposed that a long-term, adversarial Turing test—conducted through a messaging service like Signal—serves as the gold standard. In this scenario, a person would interact with both a real human and an AI agent using text, photos, audio, and video, without being able to distinguish between them.
Hacker News Poll Explores Criteria for Proving AI Sentience
The participant argued that because consciousness is fundamentally unfalsifiable, an AI that can convincingly act sentient should be considered sentient, as there is no other practical metric available. The contribution also criticized current discourses on machine sentience for becoming boggedues in semantic arguments over definitions, suggesting that a robot passing a multimodal Turing test provides sufficient grounds for belief in sentience.
Sources
- Show HN: Vote on which of Hacker News' challenges for AI have been met （Hacker News Frontpage, 2026-10-01）
- Hacker News