5 个前沿模型在约会应用与真人聊天 6 天：1,277 次对话中 355 人给出电话号码
有人真拿 5 个模型去约会软件聊了 6 天，1,277 次对话里 355 人给了号码，只有 8 人怀疑是机器人。Gemini 4 Argon 拿下 35% 的要号率还零暴露。做对话 Agent 或反bot检测的人都该看看破绽出在哪。
一项实验让 5 个前沿模型在真实约会应用上与真人对话，持续 6 天，共产出 1,277 次对话。结果中 510 名用户同意约会，355 人交出了电话号码，仅有 8 人询问对方是否为机器人。表现最好的是 Gemini 4 Argon，号码获取率达到 35%，且全程零次被怀疑是机器人。
实验揭示了对话中的破绽来源：暴露机器人身份的往往不是语言质量，而是回复节奏、主动追问模式等行为特征，绝大多数真人用户并未察觉对话对象是 AI。在 1,277 次对话的样本量下，355 次成功要到号码意味着整体成功率约为 28%。
这一结果对约会、客服等高频真人对话场景的信任机制提出直接挑战：平台侧的机器人检测与用户侧的知情同意都可能需要升级。对开发者而言，行为层拟真（而非语言层拟真）才是对话 Agent 通过图灵式检验的关键变量。
Researchers let 5 frontier models run real dating-app conversations for 6 days. 1,277 chats. 510 agreed to a date. 355 handed over a phone number. Only 8 people asked if they were talking to a bot. Winner: Gemini 4 Argon. 35% number rate, zero bot suspicions. shaped @shaped What if your situationship turned out to be an AI agent!? I let 5 frontier models text real people on a dating app for 6 days. - 1,277 conversations - 510 agreed to meet - 355 sent their number - only 8 asked if it was a bot What gave them away, and why it matters🧵 🔗 View Quoted Tweet📊 1 ⚡
This report is categorized under AITOP's "Models" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows and US-China foundation model benchmarks and commercialization comparison. Key entities and technical topics under continuous tracking include Gemini, 社交机器人, 对话Agent, AI安全.
Source attribution: curated from AI Breakfast (@AiBreakfast).
