内容来源：https://blog.google/innovation-and-ai/technology/safety-security/scams-fraud-protection/

内容总结：

谷歌公布五大反诈措施，AI技术成防骗核心防线

本周，来自政府、科技界、消费者团体和学术界的专家齐聚苏黎世，参加由谷歌安全工程中心举办的第二届欧洲、中东和非洲反诈骗峰会。针对当前网络诈骗频发、给民众造成严重经济和情感伤害的现状，谷歌宣布将通过五大举措加强用户防护，打造更安全的网络环境。

一、AI技术筑起第一道防线

在诈骗分子利用AI作恶的同时，谷歌将AI用于防御。目前，谷歌邮件系统已能阻止超过99.9%的垃圾邮件、钓鱼邮件和恶意软件，每天拦截近150亿封 unwanted 邮件。Chrome浏览器可实时预测并拦截危险网站；搜索功能每天过滤数亿垃圾页面，确保搜索结果99%无垃圾内容。2025年，谷歌系统在广告上线前便拦截了超过99%的违规广告，累计阻止或移除超过83亿条广告，其中包括6.02亿条诈骗广告。此外，手机端AI驱动的“诈骗检测”功能可实时提醒用户识别典型诈骗对话模式。

二、提供便捷安全工具

除自动防护外，谷歌推出多项易用工具帮助用户自主防范。“安全检查”功能可快速加固账户安全，支持通行密钥和两步验证。安卓系统的“圈定搜索”功能让用户长按主页键圈选可疑短信，AI将自动评估是否为诈骗并给出建议。

三、教育赋能提升防骗能力

谷歌推出互动游戏式防骗项目“Be Scam Ready”，通过模拟真实骗局帮助用户培养批判性思维。该项目已支持英语、葡萄牙语、泰语和繁体中文，今年还将新增法语、西班牙语和阿拉伯语版本。同时，谷歌.org去年承诺的500万美元资助已落地，互联网协会和牛津信息实验室正在开展基层培训，目标保护超过700万弱势群体，并培训数千名一线工作人员。

四、共享威胁数据打击诈骗源头

针对跨平台、跨国境的诈骗行为，谷歌参与创建“全球信号交换”平台，作为威胁数据的全球清算中心。该平台已存储超过12亿条信号，通过AI分析发现隐藏模式，为调查和执法提供支持。

五、政企合作摧毁犯罪网络

谷歌与英国国家打击犯罪局等执法机构直接合作，近期通过共享信号成功识别并打掉一个西非诈骗网络。此外，谷歌还签署了《在线诈骗与欺诈行业协议》，并与多家企业共享专业资源。在法律层面，谷歌已对“Lighthouse”钓鱼即服务平台和BadBox僵尸网络运营者提起诉讼，其中Lighthouse在起诉次日即被关停。

谷歌表示，打击诈骗是一项动态而持续的工作，通过将先进AI技术与广泛合作相结合，致力于为所有人构建更安全的互联网。

中文翻译：

我们对抗欺诈的五大举措：全力守护您的安全
线上欺诈极具破坏性，可能给人们带来沉重的经济和情感冲击。谷歌致力于直面这一挑战。作为努力的一部分，本周，来自政府、科技界、消费者团体和学术界的专家将齐聚苏黎世，参加由谷歌安全工程中心主办的第二次欧洲、中东及非洲反诈骗与反欺诈峰会。目标简单却雄心勃勃：加强集体行动，以瓦解当今复杂的诈骗手段。
为支持这一使命，我们在您日常使用的产品中构建了人工智能驱动的保护机制，并与行业及执法部门合作，力求领先于欺诈者。以下是我们在保护您免受诈骗和欺诈方面所做的五项工作：

1. 以人工智能工具作为第一道防线
   诈骗者利用人工智能作恶，而我们则将其用于善举。早在诈骗信息到达您之前，我们的人工智能防御系统就已开始拦截：

   • 我们阻止了超过99.9%的垃圾邮件、网络钓鱼和恶意软件进入您的Gmail收件箱，每天屏蔽近150亿封不想要的邮件。
   • 在Chrome中，我们实时预测并拦截危险网站；而在搜索中，我们每天过滤掉数亿个垃圾页面，确保您的搜索结果99%无垃圾信息。
   • 2025年，我们的系统在违规广告触达用户之前就拦截了超过99%的广告，并阻止或移除了超过83亿条广告，其中包括6.02亿条与诈骗相关的广告。
   • 设备端的人工智能驱动了Google电话应用中的“诈骗检测”功能，实时提醒用户注意典型的诈骗对话模式。

2. 为您提供安全工具
   除了自动保护机制，我们还为您提供易于使用的工具，让您掌控自身安全并即时识别诈骗。“安全检查”可帮助您通过启用通行密钥和两步验证等保护措施，快速增强账户安全。在安卓设备上，使用“圈选搜索”功能，长按主屏幕按钮并圈选可疑短信，我们的系统将利用人工智能评估其是否为诈骗信息并提供指导。（您也可以在任何手机上截屏并使用Google Lens进行相同操作。）

3. 通过教育提升防范意识
   用知识武装人们，是培养识别诈骗所需批判性思维的最有效方式之一。

   • “防骗就绪”计划是一项基于游戏的互动式项目，帮助您学习诈骗者使用的策略。通过安全模拟真实世界的诈骗场景，这种“边做边学”的方法已被证明比传统宣传活动更能有效提升防范能力。该计划目前提供英语、葡萄牙语、泰语和繁体中文版本，今年晚些时候还将推出法语、西班牙语和阿拉伯语等更多语言版本。
   • 继去年峰会上谷歌.org承诺投入500万美元用于打击欧洲和中东地区的诈骗后，受资助方国际互联网协会和牛津信息实验室正在推出基层培训计划，目标是为超过700万弱势群体提供保护，并为数千名一线工作者配备反诈骗工具。
   • 我们还定期发布关于新兴趋势的欺诈和诈骗警报，例如在线招聘诈骗或虚假包裹送达通知。

4. 共享威胁数据，从源头打击诈骗
   要打击跨平台、跨国家作案的诈骗者，我们需要协同行动。跨部门协作的最有效解决方案之一是“全球信号交换”平台，它充当威胁数据的全球信息交换中心，帮助在诈骗造成危害之前识别并打击它们。作为GSE的创始合作伙伴，谷歌既从该平台获取信息，也提供独特的威胁情报，同时我们的人工智能模型分析这些信号以发现隐藏模式，为调查和执法行动提供依据。GSE现已存储超过12亿条信号，并正在产生实际影响，助力瓦解犯罪活动。

5. 携手摧毁犯罪网络
   消除欺诈需要公共和私营部门共同采取联合方法。以下是我们正在采取的行动：

   • 我们直接与英国国家犯罪局等执法机构合作，打击欺诈活动。得益于NCA通过GSE共享的信号，我们的团队最近成功识别并瓦解了一个来自西非的欺诈网络。
   • 我们是《线上诈骗与欺诈行业协议》的签署方，该协议是线上行业参与者之间的一项约定，旨在共享集体专业知识与资源。
   • 我们采取直接法律行动，追究不法行为者的责任。我们已对大型诈骗活动提起诉讼，例如“钓鱼即服务”网络Lighthouse（我们在起诉后第二天该网络即被关闭），以及BadBox僵尸网络的运营者。

打击诈骗是一项动态且持续的努力——我们对此坚定不移。通过将先进的人工智能与广泛合作相结合，我们可以为所有人构建一个更安全的互联网。想了解更多关于我们如何保护您的信息，并获取安全上网技巧，请访问我们的安全中心。

英文来源：

Our fight against fraud: 5 ways we’re keeping you safer
Online fraud is highly disruptive, and can have a painful financial and emotional impact on people. Google is committed to tackling this challenge head-on. As part of that effort, this week, experts from across government, technology, consumer groups and academia are gathering in Zurich for the second EMEA Anti-Scams and Fraud Summit, hosted by the Google Safety Engineering Center (GSEC). The goal is simple but ambitious: Strengthen the collective action needed to disrupt today’s sophisticated scams.
To support this mission, we’re building AI-driven protections into the products you use every day and collaborating across the industry and with the authorities to stay ahead of fraudsters. Here are five ways we are working to keep you safer from scams and fraud:

1. Using AI-powered tools as the first line of defense
   While scammers are using AI for nefarious purposes, we’re using it for good. Long before a scam reaches you, our AI-powered defenses are working to block it:
   • We stop more than 99.9% of spam, phishing and malware from reaching your Gmail inbox, blocking nearly 15 billion unwanted emails every day.
   • In Chrome, we predict and block dangerous sites in real time; while in Search, we filter out hundreds of millions of spammy pages daily to ensure your results are 99% spam-free.
   • In 2025, our systems caught over 99% of policy-violating ads before they reached a person and we blocked or removed over 8.3 billion ads, including 602 million ads associated with scams.
   • On-device AI powers "Scam Detection" in Phone by Google, alerting users in real time to conversational patterns typical of scammers.
2. Enabling you with security tools
   Beyond our automated protections, we give you easy-to-use tools to control your own security and spot scams in the moment. Security Checkup helps you quickly strengthen your account security by enabling protections like Passkeys and 2-Step Verification. And with Circle to Search on Android, you can long-press your home button and circle a suspicious text message. Our systems will use AI to assess whether it’s a likely scam and provide guidance. (You can do the same on any phone by taking a screenshot and using Google Lens.)
3. Building resilience through education and awareness
   Empowering people with knowledge is one of the most effective ways to help develop the critical thinking skills to spot scams.
   • Our Be Scam Ready initiative is an interactive, game-based program that helps you learn the tactics scammers use. By safely simulating real-world scams, this “learn-by-doing” approach has been proven to build better resilience compared to traditional awareness campaigns. Be Scam Ready is available in English, Portuguese, Thai, and Traditional Chinese and is launching in more languages later this year, including French, Spanish, and Arabic.
   • Following our $5 million Google.org commitment at last year’s summit to combat scams in Europe and the Middle East, grantees Internet Society (ISOC) and Oxford Information Labs (OXIL) are rolling out grassroots training programs with the goal of protecting over 7 million vulnerable individuals and equipping thousands of frontline workers with scam resilience tools.
   • We also regularly publish fraud and scams advisories about emerging trends, such as online job scams or fake package delivery notifications.
4. Sharing threat data to disrupt scams at the source
   To stop scammers who operate across multiple platforms and countries, we need to work together. One of the most effective solutions for cross-sector collaboration is the Global Signal Exchange (GSE), which acts as a global clearinghouse for threat data helping to identify and disrupt scams, before they cause harm. As a founding partner of the GSE, Google both draws from and provides unique threat intelligence to the platform, while our AI models analyze the signals to uncover hidden patterns which inform investigations and enforcement actions. The GSE now stores over 1.2 billion signals and is having a real-world impact, helping to disrupt criminal operations.
5. Working together to take down criminal networks
   Eliminating fraud requires a joint approach that extends across the public and private sectors. Here’s what we’re doing:
   • We partner directly with law enforcement agencies, like the UK’s National Crime Agency (NCA), to disrupt fraud activity. Thanks to signals shared by the NCA through the GSE, our teams were recently able to identify and disrupt a fraud network from West Africa.
   • We are a signatory to the Industry Accord Against Online Scams and Fraud, an agreement among online industry actors to share our collective expertise and resources.
   • We take direct legal action to hold bad actors accountable. We’ve filed lawsuits against major scam operations like Lighthouse, a "Phishing-as-a-Service" network that shut down the day after we sued, and the operators of the BadBox botnet.
     Stopping scammers is a dynamic, ongoing effort – and one we are fully committed to. By combining advanced AI with broad collaboration, we can build a safer internet for everyone. To learn more about how we protect you and get tips for staying safe online, visit our Safety Center.