研究人员开发了一个名为“人类最后一道验证防线”（Humanity's Last Line of Verification, HLL）的新基准测试，用于测试多模态AI代理绕过验证码挑战的能力。该基准测试评估代理与界面进行类似人类交互的能力，而不仅仅是识别图像，并在现实条件下评估其性能。目前的前沿代理在跨越这道人类验证边界方面显示出显著的局限性，突显了在本地化、动作校准和状态跟踪方面的改进空间。 AI
影响 测试AI代理绕过人类验证系统的能力，突显了其在现实世界应用中的当前局限性。
排序理由 该集群包含一篇详细介绍AI代理新基准测试的学术论文。
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
arXiv:2606.02449v1 Announce Type: new Abstract: Multimodal agents are increasingly expected to operate interfaces on behalf of users, raising a central deployment question: can they truly substitute for humans in workflows that services deliberately protect against automation? CA…
Multimodal agents are increasingly expected to operate interfaces on behalf of users, raising a central deployment question: can they truly substitute for humans in workflows that services deliberately protect against automation? CAPTCHA verification makes this question concrete.…