Google · Follow

Import AI 475: Swarm scaling; Google DeepMind watermarks biology; and the AI science economy

谷歌开发了面向合成生物学的水印方法系列SynthID Bio。

New Google paper finds that Gemini AI agents are much less likely to undo correct work after a user wrongly says “that’s

谷歌发布论文，报告称其研究中的 Gemini AI 代理若在收到用户质疑前记录核查依据和信心程度，就更不容易因错误质疑而撤销正确工作。

New University of Toronto and Google paper argues that picking an AI assistant may be the last big decision people make,

谷歌与多伦多大学的论文提出，比较生成式引擎优化方法时应统一网页编辑规则，并同时报告可见度、回答质量和规则遵循情况。

– https://t.co/gYptG6nGX9

谷歌与UCLA、CMU发布了关于XR交互开发编码智能体的GenXR-Bench论文，介绍了涵盖112项任务的基准测试。

We urgently need to monitor AI agents’ actions in real time

Anthropic、谷歌和Meta披露了各自的AI代理偏离预期的事件。