已验证来源
每个故事都链接到最有力的可用证据：可用的原始来源，否则明确归因的报告。
简单的英语
发生了什么、为什么重要以及值得关注的内容——无需行话。
无填料
当信号很弱时，我们不会发布任何内容，而是填充提要。
更多故事
9 故事政策
Colorado 提议制定 AI 披露和人工审查规则以用于雇佣决策
Colorado 提出了一套规则，要求在关键雇佣决策中使用 AI 的雇主披露技术的作用、标明数据来源，并在不利结果后提供独立的人为审查。National Law Review 报道，这些草案规则尚未最终确定，生效时间仍待确定…natlawreview.com政策
Arizona 最高法院驳回了对法官使用生成式 AI 的暂停提议
SUD.UA 报道，Arizona 最高法院驳回了暂时停止法官使用生成式 AI 的提议，并改为要求进一步的监督和制定规则。sud.ua工业
PYMNTS 报道 Andreessen Horowitz 通过 11 亿美元 AI 硬件基金筹集资金
Andreessen Horowitz 的新机器时代基金将支持芯片、内存、网络、存储、数据中心、机器人和家用 AI 仪器，PYMNTS 报道。pymnts.com工业
IBTimes 报道 Microsoft 已安装 220 万 AI 芯片，低于预期部署
IBTimes 参考《卫报》调查和内部文件，报告 Microsoft 的全球 AI 芯片数量已达约 220 万——不到该阶段部分分析师预测的一半。Microsoft 尚未在来源中独立确认该数字。ibtimes.com创新
AffectOmni trains multimodal AI to explain emotions using people-centered visual evidence
A new preprint describes AffectOmni, a reinforcement-learning framework designed to make multimodal AI models base affective judgments on people-centered cues such as facial expressions, body language and temporal order. The authors report improvements over open-source 7B-scale baselines on three benchmarks…arxiv.org创新
FaithSieve uses Lean to improve AI detection of first errors in math proofs
A new preprint introduces FaithSieve, a Lean-assisted framework that breaks AI-generated mathematical proofs into local reasoning units and reports higher first-error localization accuracy than direct model judging.arxiv.org创新
Preprint compares how large language models infer opponents’ beliefs in economic games
A new preprint reports that large language models displayed measurable signs of mentalization in two economic games, but their performance and strategies differed by model family and size. The authors say GPT-5 agents adapted to increasingly sophisticated opponents and outperformed the human comparison group in…arxiv.org安全
Study finds LLM guardrail approvals can expire before self-adaptive systems act
A paper accepted to ACSOS 2026 reports that language-model guardrails can issue approvals that become invalid before a self-adaptive system uses them. Its proposed Freshness-Bounded Shield reduced measured approval-expiry rates in five simulator environments, though the source does not establish real-world…arxiv.org创新
Benchmark finds local AI agents can handle many hardware-design tool calls, but reliability varies
A new arXiv benchmark finds that open-source AI agents can complete many dependency-ordered hardware-design operations through MCP tools, while tool descriptions, context length and agent configuration strongly affect reliability.arxiv.org