内容来源：https://www.wired.com/story/how-to-opt-out-of-google-search-new-ai-data-training/

内容总结：

谷歌默认启用新设置：用户搜索数据将被用于AI训练，需手动关闭

近日，谷歌向用户发送一封题为“搜索服务新隐私设置”的邮件，宣布将在未来数月内全球逐步推行一项新政策：调整搜索历史数据的处理方式。这意味着，用户通过谷歌搜索上传的图片、语音搜索记录、谷歌翻译音频等内容，都有可能被保存至账户，并用于改进谷歌的AI模型。

这项名为“搜索服务历史记录”的新选项，在用户首次访问设置页面时已默认开启。如果用户此前曾关闭“网页与应用活动”或“搜索个性化”功能，则此选项可能处于关闭状态。此外，用于AI训练的“保存媒体文件”选项也已被默认勾选。用户若想关闭该功能，需前往谷歌“我的活动”页面，选择“搜索服务历史记录”标签，手动取消勾选“保存媒体文件”并删除已有活动记录。

值得注意的是，即便用户后期删除原始活动数据，已被用于AI训练的媒体文件仍将与谷歌账户解绑，并保留长达4年。谷歌发言人戴维斯·汤普森回应称，“新设置有助于用户获得更精准的搜索结果，并可随时开启或关闭。”但他未就默认开启一事做出进一步说明。

电子前沿基金会高级安全与隐私活动家索林·克洛索夫斯基指出，谷歌凭借其庞大的多平台用户基础，在数据收集领域拥有独特优势。用户长期依赖谷歌服务，对数据收集已习以为常，即使对政策不满，也未必会转向其他替代平台。他批评谷歌将选择权交给用户的做法，认为“默认开启、用户选择退出”的方式远不及“用户主动选择加入”更为透明和公平。

美国消费者联合会AI与隐私主管本·温特斯则认为，此类设置正加剧用户的疲惫感甚至虚无主义——“用户越来越感到无能为力，因为每一个细微的数据都可能被榨取。”他提醒，无论是搜索还是其他日常服务，用户应主动检查隐私设置，避免数据被无意用于AI训练而不自知。

中文翻译：

每当我收到邮件，讲述另一家公司计划如何以日益侵入的方式将个人数据用于AI训练时，我灵魂的一小部分便会枯萎。最近我收到了谷歌的一封此类邮件，主题是：“搜索服务的新隐私设置”。这是谷歌在未来几个月内全球推行的一部分，将改变其处理用户搜索历史数据的方式。

每一种媒体内容——从你上传用于反向图片搜索的照片，到你与谷歌翻译对话的音频——都可能会保存在你的账户中，并用于改进谷歌的AI模型。

这个名为“搜索服务历史记录”的谷歌账户设置新选项，在我第一次访问该页面时已经处于启用状态。（如果用户之前关闭了谷歌的“网页和应用活动”以及“搜索个性化”开关，那么它会处于关闭状态。）此外，用于保存我所有从谷歌搜索上传的媒体内容以进行AI训练的复选框，也已经勾选上了。真棒。

当这项设置推送到你的账户时，你可以访问谷歌的“我的活动”页面，然后选择“搜索服务历史记录”选项卡来退出。该页面会让你清楚了解谷歌从你的搜索历史中保存了哪些内容。你也可以在这里关闭整个设置并删除你的活动记录。如果你不希望你的图片上传被用于AI训练，务必取消勾选“保存媒体文件”旁边的复选框。

现在进行这一更改比以后再做要值得。一旦你的媒体数据被投入AI搅拌机，你能做的就很有限了。“如果你保存的媒体内容被用于训练我们的AI模型，它将与你的谷歌账户断开连接。即使你删除了原始活动记录，这些训练数据也将被保留最多4年，”在我关闭此功能时弹出的窗口如此写道。我随机上传的图片搜索内容在数字空间中飘荡如此之久，真是漫长。

“这些新设置帮助用户获得更相关的结果，并重温他们的搜索记录——包括视觉搜索和语音搜索——而且可以随时开启或关闭，”谷歌发言人戴维斯·汤普森在通过电子邮件回应置评请求时表示。他没有回答《连线》杂志关于此功能默认开启的问题。

这里真正保存的是什么？谷歌明确指出，它开始存储的远不止你在搜索框中输入的内容。“你保存的媒体内容包括你在与搜索服务互动时产生的图像、文件、以及音频和视频记录，”该页面的描述写道。“这包括谷歌镜头图像、搜索直播或翻译练习的录音、你上传的内容以及语音搜索。”

这是因为AI模型不仅需要文本数据来改进；它们还需要其他形式的多样化输入，比如音频或视频。如果谷歌能够从其庞大的用户群中收集更多数据，以及更多类型的数据，那么它或许能比竞争对手创新得更快。

谷歌庞大的用户群分布在多项服务中，这使其在数据收集方面具有优势。“与许多其他公司相比，谷歌处在一个独特的位置，”电子前哨基金会的高级安全与隐私活动家索林·克洛索夫斯基表示。“因为他们提供如此多的服务，人们已经使用了很长时间，并且对收集的数据量已经变得相当习惯和自满。”人们日常使用的应用具有一种内置的惯性，因此用户不喜欢的改变可能不足以促使他们转向替代服务。

被迫选择退出AI训练似乎已成为各网站和平台的常态。但本不必如此。“我认为‘选择加入’实际上是对这些公司的底线要求，”克洛索夫斯基说。“要求用户有意识地选择启用这些功能，是他们至少能做到的。”克洛索夫斯基表示，如果这些功能不是默认自动开启，谷歌就必须向用户更有力地说明，为什么这些功能可能是有用的。

在谷歌于6月23日发送到我的测试账户的邮件中，第一句话将这一改变描述为让我“对保存的历史记录拥有更多控制权”。随后谷歌在邮件中举例说明了保存这些媒体内容可能带来的便利。“例如，这让你可以重温你过去使用镜头进行的视觉搜索，或者继续关于你听过的一首歌的搜索直播对话。”相比之下，值得注意的是，谷歌在邮件末尾附近说明这些保存的媒体内容将用于AI模型训练后，并未提供类似例子。相反，邮件只是继续叙述下一个细节。

这是另一个值得日常用户放缓节奏来处理的重大软件变更。“这给消费者增加了一层额外的计算，他们需要判断自己是否还愿意继续使用这个已经用了很久的工具，”美国消费者联合会AI与隐私事务主任本·温特斯表示。

我常常感到不堪重负，因为我总是需要为每一项服务选择退出数据训练。这让我感觉自己像个傻瓜，无论怎么做，都可能错过埋藏在这些设置中的某些东西。

温特斯认为，谷歌的这一改变是将避免AI训练的责任推给了用户，这可能导致广大用户近乎虚无主义的疲惫感。“人们越来越感到无能为力和绝望，甚至连保护自己数据的尝试都放弃了，因为你的一丁点东西都会被榨干，”他说。

英文来源：

A little piece of my soul shrivels up every time I get a message laying out how another company plans to use personal data in ever encroaching ways for AI training. I got one of those emails recently from Google, with the subject line: “New privacy settings for Search services.” It’s part of Google’s global rollout happening over the next few months that will change how it handles users' Search history data.
Every piece of media, from photos you upload for reverse image searches to audio of you speaking with Google Translate, may be retained in your account and used to improve Google’s AI models.
This new option in Google’s account settings, called Search Services History, was already enabled when I visited the page for the first time. (If a user previously disabled Google’s Web & App Activity and Search Personalization toggles, then it would be off.) Also, the box to save all my uploaded media from Google Search for AI training was already checked. Great.
When this rolls out to your account, you can visit Google’s My Activity page and then select the Search Services History tab to opt out. This page gives you a solid sense of what Google saves from your Search history. It’s also where you can turn off the entire setting and delete your activity. It’s critical to uncheck the box next to Save media if you don’t want your image uploads used for AI training.
It’s worth doing this change now rather than later. There’s not much you can do after your media data is thrown into the AI blender. “If your saved media is used to train our AI models, it is disconnected from your Google Account. This training data will be kept for up to 4 years, even if you delete the original activity,” read the pop-up when I turned off this feature. That’s such a long time for my random image Search uploads to float around in the digital ether.
“These new settings help users get more relevant results and revisit their searches—including visual and voice searches—and they can be turned on or off at any time,” says Davis Thompson, a Google spokesperson, over email when reached for comment. He did not answer WIRED’s question about this feature being on by default.
What’s really being saved here? Google is clear that it's starting to store much more than just what you type into Search. “Your saved media includes your images, files, and audio and video recordings from your interactions with Search services,” reads the page’s description. “This includes things like Google Lens images, recordings from Search Live or Translate speaking practice, content you upload, and voice searches.”
This is because AI models don’t just need text data to improve; they need diverse inputs in other forms, like audio or video. If Google can gather more data, and more types of data, from its vast array of users, then maybe it could innovate faster than its competitors.
Google's massive user base is spread across multiple services, giving it an edge in data collection. “Google is in a unique spot compared to a lot of the other companies with this,” says Thorin Klosowski, a senior security and privacy activist at the Electronic Frontier Foundation. “Because they offer so many services that people have been using for so long and have grown pretty comfortable and complacent with the amount of data collected.” Apps that people use every day have a kind of built-in inertia, so changes that users don't like may not be enough to drive them to alternative services.
Being forced to opt out of AI training seems to be the standard across sites and platforms. It doesn’t have to be this way. “I think ‘opt in’ is really asking the bare minimum of these companies,” Klosowski says. “Asking their users to consciously choose to enable these features is the least they can do.” Google would have to make a stronger case to users on why these features could be helpful if they weren’t automatically just turned on, Klosowski says.
In Google’s email sent to my testing account on June 23, the first sentence framed this change as giving me “even more control over saved history.” Google then provided examples in the message showing how saving this media may be helpful. “For example, this lets you revisit your past visual searches with Lens or continue a Search Live conversation about a song you heard.” In contrast, it’s notable that Google didn’t provide similar examples after stating near the end of the email that this saved media will be used for AI model training. Rather, the message just continued to the next detail.
This is another major software change that is worth slowing down to process the change for everyday users. “It creates this extra layer of math that a consumer has to do about whether they feel comfortable using the tool they've been using for a long time,” says Ben Winters, director of AI and privacy at the Consumer Federation of America.
I’m constantly overwhelmed that it's always on me to opt out of data training for every service. It leaves me feeling like a schmuck who’s probably going to miss something buried in all these settings, no matter what.
Winters sees this change from Google as placing the onus on users to avoid AI training, which may contribute to widespread user exhaustion bordering on nihilism. “There’s an increasing feeling of powerlessness and hopelessness about even trying to protect your data, because every little thing is going to be squeezed out of you,” he says.