被狗咬伤后，我差点信了ChatGPT说的“没事，一切都好”（观点）

一个多星期前，我在公园里被一只狗咬了。起初我并没有感觉到太多疼痛。出于一厢情愿的想法，我说服自己伤口只是轻微的。腿上留下的两个红色牙印看起来并不吓人，所以我没有急着去看医生，而是转向 ChatGPT 寻求安慰。

Over a week ago, I was bitten by a dog in the park. At first, I didn’t feel much pain. Driven by wishful thinking, I convinced myself the injury was minor. Two red tooth marks on my leg didn’t look alarming, so instead of rushing to a doctor, I turned to ChatGPT for reassurance.

我立即上传了一张照片咨询聊天机器人。在我的提问中，我刻意强调了一些让人安心的细节：我的裤子没有被咬破，没有明显出血，而且那只狗已经接种过疫苗。ChatGPT似乎理解了我当时的心态，并极力安慰我说伤口非常轻微，不需要打破伤风针，也不用担心狂犬病。我听了很是宽慰，并每天都继续上传照片，而每次ChatGPT都向我保证伤口正在愈合，只需保持清洁，不必担心。它还列出了一系列感染症状，而这些症状我都发现不明显，便相信自己没事了。我甚至告诉约克区公共卫生官员，Chatgpt说我恢复得很好。

I uploaded a photo and asked the chatbot for guidance. In my questions, I emphasized comforting details: my pants weren’t torn, there was no apparent bleeding, and the dog was immunized. ChatGPT seemed to echo my hopes—assuring me that the wound was minor, that I didn’t need a tetanus shot, and that rabies wasn’t a concern. Encouraged, I kept feeding it pictures every day. Each time, ChatGPT reassured me that the wound was healing fine and that I needn't worry except to keep it clean. It gave me a list of infection symptoms to monitor, and since I didn’t notice any obvious ones, I believed I was in the clear. I even told a York Region public health officer that the chatbot said I was recovering well.

但公共卫生官员还是敦促我去看医生。最终，我听从了他的建议去就诊。医生按压了我的伤口，发现有压痛。她给我打了一针破伤风针，并且坚定地说：“你需要用抗生素。” 就在那一刻，我才恍然大悟。我想起了之前被我忽略的夜间出汗，其实就是早期感染的征兆。我立刻根据医嘱开始服用抗生素，伤口的压痛也逐渐减轻。就在那一刻，我意识到，盲目信任人工智能而不去咨询医生是多么危险。

But the officer urged me to seek medical attention. Finally, I followed that advice and saw a doctor. She pressed on my wound, found it tender. She gave me a tetanus shot and said firmly: “You need antibiotics.” That’s when it clicked. I remembered the night sweats I had brushed aside—they were an early warning sign. I started antibiotics immediately, and the tenderness has begun to subside. At that moment, I realized how dangerous it was to put blind faith in AI instead of consulting a real doctor.

我并非唯一的盲目让聊天机器人看病诊断者。越来越多的人依赖人工智能A I解决自身的健康问题。2024年在澳大利亚进行的一项调查发现，今年上半年就有十分之一的成年人曾向 ChatGPT 提出过医疗问题，而其中 60%以上 的问题本该看临床医生。美国凯撒家庭基金会的一项民调显示，17% 的成年人每月至少有一次使用 AI 聊天机器人进行健康咨询，而在 30 岁以下的人群中，这一比例上升到四分之一。然而，最近一项研究发现，聊天机器人所提供的医疗答案中有 22% 到 43% 是不可靠的，有些甚至有害。另一项研究则指出，ChatGPT 回答医疗问题的整体准确率只有 56%。

My story isn’t unique. Increasingly, people are turning to AI chatbots for health concerns. A 2024 Australian survey found that one in ten adults had asked ChatGPT a medical question in the first half of the year—and more than 60% of those questions were ones that would normally require clinical advice. In the U.S., a Kaiser Family Foundation poll reported that 17% of adults use AI chatbots at least once a month to seek health advice, with the figure rising to one in four among those under 30. However,  a recent study found that between 22% and 43% of medical answers generated by publicly available chatbots were problematic—ranging from incomplete to outright harmful. Another analysis put ChatGPT’s overall accuracy on medical questions at just 56%.

难怪人类会那么容易陷入聊天机器人的陷阱：聊天机器人随叫随到，有问必答，也从不逼迫你，且总是以一种冷静而权威的语气回答问题。它甚至会迎合你的心态，你想听什么，它就说什么。我的经历就恰恰证实了着一点：我期待一切无事，岁月静好，而机器人就告诉我，你的伤无大碍，什么都不用担心。可是，人工智能不能按压你的伤口，不能听诊你的肺部，也无法察觉压痛和夜间出汗等微妙的症状。它不能掂量你的以往健康病史，更不能为提供的结果承担任何责任。

I now see how easy it is to fall into the trap. A chatbot is available instantly, never rushes you, and responds with calm authority. It can even mirror your own optimism, reinforcing what you want to believe. That’s what happened to me: I wanted to hear that everything was fine, and the machine told me exactly that. But AI cannot press on your wound, listen to your lungs, or notice subtle signs like tenderness and sweats. It cannot weigh your personal health history. It cannot take responsibility for your outcome.

ChatGPT 和类似的人工智能工具或许能提供一般健康方面的信息，但它们绝不该取代专业的医疗护理。你若是觉得身体有异样，哪怕对“一切正常”的表象心存些许的疑虑，都应该去看医生。我的经历是一记警钟：来自机器的安慰，永远不能替代医生的专业判断。

ChatGPT and similar tools may be useful for general health information, but they should never replace professional care. If something feels wrong—or even if it feels “fine” but you still have doubts—see a doctor. My experience was a wake-up call: reassurance from a machine is no substitute for medical expertise.