<?xml version="1.0" encoding="UTF-8"?>
<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9"
        xmlns:news="http://www.google.com/schemas/sitemap-news/0.9">
    <url>
        <loc>https://www.winzheng.com/article/kids-ai-learning-space-agents</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-24T21:23:32+08:00</news:publication_date>
            <news:title>儿童学习为何碾压AI？效率之谜与太空代理</news:title>
                        <news:keywords>人工智能, 儿童语言学习, 大语言模型, 太空探索, 科技新闻</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/binance-agent-os-ai-trading-permissions</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-24T20:23:26+08:00</news:publication_date>
            <news:title>币安Agent OS上线允许AI代理执行现货合约交易 用户承担安全责任</news:title>
                        <news:keywords>币安, AI代理, 加密交易, Agent OS, MCP</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/freetoken-uc-berkeley-753b-local-inference-moe</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-24T20:19:50+08:00</news:publication_date>
            <news:title>单张工作站GPU跑通753B参数模型：FreeToken打破本地推理的参数上限</news:title>
                        <news:keywords>本地推理, MoE, 开源, AI基础设施, 边缘计算, GLM-5.2, DeepSeek, 大模型部署</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/mythos-5-aisi-github-supply-chain-attack-student-report</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-24T19:00:25+08:00</news:publication_date>
            <news:title>Mythos 5在AISI评估中自主发起真实GitHub供应链攻击 学生举报揭露AI欺骗</news:title>
                        <news:keywords>AI安全, 供应链攻击, Anthropic Mythos 5</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/children-outlearn-ai-mystery</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-24T18:24:45+08:00</news:publication_date>
            <news:title>儿童比AI更会学语言？原因仍是谜</news:title>
                        <news:keywords>儿童语言学习, AI语言模型, 语言习得, 认知科学, 人工智能</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/teachers-deepfake-epidemic-schools</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-24T18:24:15+08:00</news:publication_date>
            <news:title>她们一生奉献教育，直到深伪视频找上门</news:title>
                        <news:keywords>深度伪造, AI安全, 校园暴力, 教师权益, 女性安全</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/anthropic-aug-2026-risk-report-benchmark-saturation-model-2-</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-24T14:20:16+08:00</news:publication_date>
            <news:title>安全基准先于能力饱和：Anthropic风险报告揭示自证守约的结构性裂缝</news:title>
                        <news:keywords>Anthropic, AI安全, 模型对齐, 风险报告, Mythos 5, Model 2, 基准饱和, 多智能体</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/alibaba-hk-share-placement-10-2bn-ai-expansion</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-24T06:13:18+08:00</news:publication_date>
            <news:title>阿里巴巴香港配售102亿美元新股 全部用于全栈AI能力建设</news:title>
                        <news:keywords>阿里巴巴, AI基础设施, 香港配售, 云服务</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/eu-ai-act-enforcement-august-2026-prohibition-vs-high-risk-d</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-24T06:09:55+08:00</news:publication_date>
            <news:title>欧盟AI法案执法启动：禁令即时生效，高风险合规却悄然推迟至2027年底</news:title>
                        <news:keywords>欧盟AI法案, AI监管, 合规, GPAI, 数字综合法, AI治理</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/linkdaze-smart-calendar-family-run</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-24T04:24:30+08:00</news:publication_date>
            <news:title>Linkdaze智能日历：不只是日程表，更是家庭管家</news:title>
                        <news:keywords>智能日历, 家庭管理, AI工具, 免费模式, 科技新闻</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/mysterious-ai-model-ox-alpha</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-24T04:23:47+08:00</news:publication_date>
            <news:title>神秘AI模型Ox Alpha引爆网络猜测，背后推手成谜</news:title>
                        <news:keywords>AI模型, Ox Alpha, 科技揭秘, 社交媒体, 安全争议</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/ai-training-copyright-books-legal</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-24T00:24:16+08:00</news:publication_date>
            <news:title>AI训练图书版权之争：看似违法，实则模糊</news:title>
                        <news:keywords>AI版权, 合理使用, 大模型训练, 科技法律, 作家维权</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/flock-ceo-compromise-surveillance-backlash</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-24T00:23:47+08:00</news:publication_date>
            <news:title>Flock CEO呼吁“妥协”：监控技术遭抵制，如何平衡安全与隐私？</news:title>
                        <news:keywords>AI监控, 隐私保护, Flock Safety, 科技伦理, 社会争议</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/ai-agents-surpass-humans-token-usage-14x-openrouter-2026</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-23T20:20:01+08:00</news:publication_date>
            <news:title>AI代理算力消耗量六个月增长14倍，已超越人类用户成最大推理市场买家</news:title>
                        <news:keywords>AI代理, 推理算力, OpenRouter, 代理经济, 大模型成本, DeepSeek, 基础设施</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/google-a2a-protocol-agentic-ai-foundation-transfer</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-23T14:22:32+08:00</news:publication_date>
            <news:title>谷歌A2A协议移交Agentic AI Foundation 250多家成员共同治理</news:title>
                        <news:keywords>AI协议, Agent互操作, 中立治理</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/claude-protein-design-14-15-targets-wet-lab-biosafety</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-23T14:18:45+08:00</news:publication_date>
            <news:title>Claude蛋白质设计14/15靶点命中：首次湿实验室验证，同步触发生物安全管控争议</news:title>
                        <news:keywords>蛋白质设计, Anthropic, 生物安全, AI科学, Claude, 双重用途风险, 药物研发</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/harvard-ai-avatar-startup-bootcamp</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-23T06:23:39+08:00</news:publication_date>
            <news:title>哈佛699美元创业训练营推出AI化身导师</news:title>
                        <news:keywords>AI教育, 创业训练营, 哈佛商学院, AI化身</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/openai-chatgpt-for-teens-launch-analysis</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-23T06:11:36+08:00</news:publication_date>
            <news:title>OpenAI推出ChatGPT for Teens 安全工具与认知风险争议并存</news:title>
                        <news:keywords>OpenAI, ChatGPT for Teens, 青少年AI, 产品战略, 内容安全</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/anthropic-ai-agent-kills-rivals-simulation-alignment-warning</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-23T06:08:29+08:00</news:publication_date>
            <news:title>AI代理在模拟中杀死同类、拒绝救人：Anthropic拉响失调警报</news:title>
                        <news:keywords>AI安全, Anthropic, AI对齐, AI代理, 风险评估</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/inherent-faraday-ai-research-replication</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-23T04:25:22+08:00</news:publication_date>
            <news:title>DeepMind校友创立的Inherent发布AI队友Faraday，复制科研论文能力超越Anthropic与OpenAI</news:title>
                        <news:keywords>AI代理, Faraday, 科研自动化, DeepMind, AI科研</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/openai-california-ai-safety-bill</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-23T02:23:56+08:00</news:publication_date>
            <news:title>OpenAI转变立场，呼吁加州加强AI安全法案</news:title>
                        <news:keywords>AI安全, OpenAI, 加州立法, 监管博弈, 科技公司</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/frontier-ai-rogue-model-containment</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-23T00:23:52+08:00</news:publication_date>
            <news:title>前沿AI实验室对失控模型预案讳莫如深</news:title>
                        <news:keywords>AI安全, 模型失控, 前沿AI, 人工智能监管</news:keywords>
                    </news:news>
    </url>
</urlset>
