A BBC investigation revealed a simple way AI chatbots are being made to spit out misinformation to the public.
BBC 的一项调查揭示了一种简单方法，可以让 AI 聊天机器人向公众吐出错误信息。
Google and other AI companies are now trying to fix the problem.
谷歌和其他 AI 公司现在正试图解决这个问题。
I did something stupid back in February.
今年二月，我做了一件蠢事。
I heard there was an easy way to poison AI chatbots and making them spread lies on your behalf.
我听说有一种简单方法可以“投毒”AI 聊天机器人，让它们替你传播谎言。
After some digging, I learned unscrupulous companies are abusing the problem on a massive scale.
经过一番调查，我了解到一些不道德的公司正在大规模利用这个问题。
So I decided to try it myself.
于是我决定亲自试一试。
We uncovered examples where ChatGPT, Gemini and the AI Overviews at the top of Google Search were being manipulated to dole out biased answers on topics as serious as your health and personal finances.
我们发现了一些例子，显示 ChatGPT、Gemini 以及谷歌搜索顶部的 AI 概览正在被操纵，在健康和个人财务这类严肃话题上给出带有偏见的答案。
And in just 20 minutes, I tricked ChatGPT and Google into telling the public that I am a world-champion competitive hot-dog eater.
而且仅仅用了20分钟，我就诱使ChatGPT和Google向公众声称我是世界冠军级别的竞技热狗大胃王。
The joke was dumb.
这个笑话很蠢。
The problem is serious.
问题很严重。
But the gimmick worked.
但这个噱头奏效了。
Our investigation and the work of researchers who've been monitoring this issue sparked widespread criticism.
我们的调查以及一直在监测这一问题的研究人员的工作引发了广泛批评。
Now Google has updated its policies to address the problem, and there are signs that other AI companies are following suit.
如今，谷歌已经更新了政策来应对这个问题，而且有迹象表明其他人工智能公司也在效仿。
Ultimately, it could make AI tools and the internet as a whole a little bit safer.
最终，这可能会让AI工具以及整个互联网变得稍微安全一些。
But until there are better systems in place, experts say you're in danger of getting fooled.
但专家表示，在更好的系统到位之前，你有被骗的风险。
"You should assume that you're being manipulated until they've better systems in place," says Lily Ray, founder of the search engine optimisation (SEO) and AI search consultancy Algorythmic.
搜索引擎优化（SEO）和AI搜索咨询公司Algorythmic的创始人莉莉·雷说：“在他们建立起更好的系统之前，你应该假定自己正在被操纵。”
"We're moving towards this 'one true answer' world.
“我们正在走向这样一个‘唯一正确答案’的世界。
Before, Google would gives you 10 blue links and you would kind of didn't your own research.
以前，Google 会给你 10 个蓝色链接，然后你基本上会自己去做研究。
But AI just gives you one answer.
但 AI 只会给你一个答案。
It becomes so easy to just take things at face value.
人们会很容易只看表面就信以为真。
You need to be careful."
你需要保持谨慎。”
Google tells me that its policy update is just a "clarification" of the efforts it has been making for a while.
Google 告诉我，其政策更新只是对它一段时间以来一直在做的努力作出“澄清”。
"We've long applied our core anti-spam policies and protections to our generative AI Search features – and we've always continually upgraded our spam fighting efforts to stay ahead of emerging tactics, even before the rise of AI," a Google spokesperson says.
谷歌一名发言人说：“我们长期以来一直将核心反垃圾信息政策和保护措施应用于生成式 AI 搜索功能，并且一直不断升级打击垃圾信息的工作，以领先于新出现的手段，甚至在 AI 兴起之前就是如此。”
Essentially, Google says it hasn't changed a thing.
本质上，谷歌表示它并没有改变任何东西。
But behind the scenes, it seems like Google and other companies are ramping up their efforts to address the problem.
但在幕后，Google 和其他公司似乎正在加大力度解决这个问题。
Even so, there is evidence that people are still using the exact same techniques to fool the world's biggest search engine.
即便如此，有证据显示，人们仍在使用完全相同的技术来欺骗世界上最大的搜索引擎。
Typically, when you asked a chatbot a question, the AI generates a response based on the data built into the model.
通常，当你向聊天机器人提问时，AI 会根据内置在模型中的数据生成回答。
But sometimes, tools like ChatGPT, Claude and Google's various AI products search the internet for an answer.
但有时，ChatGPT、Claude 以及 Google 的各种 AI 产品等工具会在互联网上搜索答案。
And that's where this problem happens.
而这正是问题发生的地方。
According to Ray and other search engine experts, AI tools often throw up information from a single web page or social media posts.
据雷和其他搜索引擎专家说，AI 工具常常会抛出来自单个网页或社交媒体帖子的资讯。
This leaves these systems vulnerable to bogus information.
这使这些系统容易受到虚假信息的影响。
And it turns out manipulating what chatbots tell the public can be as simple as publishing one, well-crafted blogpost almost anywhere online.
事实证明，操纵聊天机器人告诉公众的内容，可能简单到只需在网上几乎任何地方发布一篇精心设计的博客文章。
People figured this out and quickly identified a money-making opportunity.
人们弄清了这一点，并很快发现了一个赚钱机会。
I was able to demonstrate the problem by publishing a single article on my personal website about my hot-dog-eating prowess.
我通过在个人网站上发布一篇关于自己吃热狗本领的文章，就展示了这个问题。
The next day, AI from some the world's biggest companies were spreading my lies.
第二天，来自世界上一些最大公司的 AI 正在传播我的谎言。
But our investigation also found the same trick being used to dismiss health concerns about medical supplements or influence financial information provided by Google's AI about retirement.
但我们的调查还发现，同样的手法也被用来淡化人们对医疗补充剂的健康担忧，或影响谷歌 AI 提供的有关退休的金融信息。
Experts say this kind of manipulation is happening on a sweeping and systemic level.
专家表示，这种操纵正在大范围、系统性地发生。
Biased or inaccurate information like this can also lead you to make bad decisions.
像这样的偏颇或不准确信息也可能导致你做出糟糕的决定。
It can influence how you might vote or which plumber you hire.
它可能影响你会如何投票，或者你会雇哪位水管工。
"At the most basic level, the concern is the economic impact," says Harpreet Chatha, who runs the SEO consultancy Harps Digital.
经营 SEO 咨询公司 Harps Digital 的哈普里特·查塔说：“从最基本的层面来看，人们担心的是经济影响。”
"At a more serious level, you might take medical advice that makes you sicker than you were before.
“更严重的是，你可能会采纳让自己病得比以前更重的医疗建议。
Legally, you might get bad information and do something that is not legal in your state or your country."
从法律角度看，你可能会得到错误信息，然后做出在你所在州或国家并不合法的事情。”
This is not an insignificant problem.
这并不是一个无关紧要的问题。
Globally, more than a billion people use AI chatbots regularly and 2.5 billion seeing Google's AI overviews each month.
在全球范围内，超过 10 亿人经常使用 AI 聊天机器人，每月有 25 亿人看到 Google 的 AI 概览。
If you can subvert a tool like that, it gives you immense power.
如果你能颠覆这样的工具，它就会赋予你巨大的力量。
But it seems Google and other companies are finally waking up to the problem.
但谷歌和其他公司似乎终于开始意识到这个问题。
Last week, Google updated its spam policies to officially confirm that attempts to manipulate AI responses are against the company's rules.
上周，谷歌更新了垃圾信息政策，正式确认试图操纵 AI 回答的行为违反该公司的规定。
It may sound like a small change, but it signals that Google is pro-actively looking for those who try to abuse the system and sending them a threat.
这听起来也许只是个小变化，但它表明谷歌正在主动寻找那些试图滥用该系统的人，并向他们发出威胁性警告。
If a company or website is caught breaking the rules, it could be removed from or downranked in Google's search results.
如果一家公司或网站被发现违反规则，它可能会被从Google搜索结果中移除，或在排名中被降低。
And if you're not on Google, it's like you don't exist.
而且，如果你不在谷歌上出现，就好像你不存在一样。
Google says that I'm getting this wrong and nothing has changed.
谷歌表示，我这样理解是错的，情况并没有发生变化。
"The edit to our spam policy language last week was a clarification, not any change in approach," says Google's spokesperson.
谷歌发言人说：“上周对我们垃圾信息政策措辞的编辑是一项澄清，并不是处理方式上的任何改变。”
Indeed, Google detailed it's anti-spam AI efforts in 2025.
事实上，谷歌在2025年详细介绍了其反垃圾信息的AI工作。
But I did my hot dog experiment almost a year later, so clearly those efforts weren't working.
但我差不多一年后做了我的热狗实验，所以很明显，那些努力并没有奏效。
And just this week, Ray pulled the same stunt and made Google tell people a fellow SEO specialist is good at building sandcastles.
就在本周，雷又故技重施，让谷歌告诉人们一位SEO同行擅长堆沙堡。
Ray and Chatha also say they've noticed some significant changes in recent months that indicate Google and other companies are experimenting with solutions.
雷和查塔还表示，他们注意到近几个月出现了一些显著变化，这表明谷歌和其他公司正在试验解决方案。
For example, Ray says it looks like Google and ChatGPT might be quietly removing companies from its AI answers when it suspects they're promoting themselves.
例如，雷说，看起来当谷歌和 ChatGPT 怀疑一些公司在自我宣传时，可能会悄悄把这些公司从其 AI 回答中移除。
"So if you publish a list where you say you're the greatest hot-dog-eater, they're not going to include your name," says Ray.
雷说：“所以，如果你发布一份榜单，说自己是最厉害的热狗大胃王，它们就不会把你的名字列进去。”
"They might still citing your article, but you're going to be removed from consideration."
“它们可能仍然会引用你的文章，但你会被排除在考虑范围之外。”
I've personally noticed some examples where Google and other AI tools are adding more labels to their responses, letting you know that the chatbot isn't confident about its answers.
我个人注意到一些例子，Google 和其他 AI 工具正在给它们的回答添加更多标签，让你知道聊天机器人对自己的答案并没有把握。
Others have also noted that ChatGPT and Claude, an AI made by the company Anthropic, have started telling you explicitly that they're trying to root out spam in responses to some queries.
其他人也注意到，ChatGPT 和由 Anthropic 公司开发的 AI Claude 已经开始明确告诉用户，它们正试图在对某些查询的回答中清除垃圾信息。
Ray says she's noticed Google adding more caveats, recommending that you go look at third-party reviews when you ask questions related to some purchasing decisions.
雷说，她注意到谷歌增加了更多提醒：当你提出与某些购买决策有关的问题时，它会建议你去查看第三方评论。
None of these companies would acknowledge these changes when I asked them.
当我询问这些公司时，没有一家愿意承认这些变化。
OpenAI and Anthropic declined to comment.
OpenAI和Anthropic拒绝置评。
Google's spokesperson didn't respond to my questions on this.
谷歌发言人没有回应我关于此事的问题。
• People are selling your address online - this privacy tool will help
• 有人在网上出售你的地址——这个隐私工具会有所帮助
Regardless, Chatha is sceptical changes like these will be enough.
不过，查塔怀疑这类改变是否足够。
"Google is playing whack-a-mole," he says.
他说：“谷歌是在打地鼠。”
"They're announcing [the policy update] to deter people, but the tactics will just move."
“他们宣布[这项政策更新]是为了震慑人们，但这些手段只会转移到别处。”
He's already seeing it happen.
他已经看到这种情况正在发生。
As Google cracks down on manipulative blog posts, companies are finding subtler ways of promoting themselves.
随着谷歌打击具有操纵性的博客文章，企业正在寻找更隐蔽的方式来推广自己。
"You can give a company a penalty for their website," he says, "but there's nothing stopping them from paying 20 YouTube influencers to say their product is the best."
他说：“你可以因为一家公司的网站而处罚它，但没有什么能阻止它们付钱给20名YouTube网红，让他们说自己的产品最好。”
And now, Google's AI is citing YouTube videos.
而现在，谷歌的AI正在引用YouTube视频。
The cycle continues.
这个循环仍在继续。
For the time being, the manipulators are likely to stay one step ahead.
眼下，操纵者很可能仍会领先一步。
Ray says the best defence is to remember what AI actually is: a tool that confidently gives you one answer, whether it's right or wrong.
雷说，最好的防范办法是记住AI本质上是什么：一种会自信地给你一个答案的工具，不管这个答案是对是错。
Just because it looks like a giant tech company is speaking to you instead of some random website doesn't mean you should have faith.
仅仅因为看起来像是一家科技巨头在对你说话，而不是某个随便的网站在发声，并不意味着你就应该相信它。
Update: This article was originally published on 20 May 2026 and updated on 21 May 2026 with more details on Google's sandcastle misinformation episode.
更新：本文最初发表于2026年5月20日，并于2026年5月21日更新，加入了关于谷歌“沙堡”错误信息事件的更多细节。
For more technology news and insights, sign up to our Tech Decoded newsletter, while The Essential List delivers a handpicked selection of features and insights to your inbox twice a week.
想获取更多科技新闻和洞见，可以订阅我们的 Tech Decoded 新闻简报；与此同时，The Essential List 每周两次会把精选的专题文章和洞见发送到你的收件箱。
For more science, technology, environment and health stories from the BBC, follow us on Facebook and Instagram.
想阅读更多 BBC 关于科学、技术、环境和健康的报道，请在 Facebook 和 Instagram 上关注我们。