A study has found that using artificial intelligence chatbots for medical advice can be dangerous. A significant portion of AI responses could lead to serious health consequences for patients.
The dangers of AI medicine
Study results
Scientists analyzed the effectiveness of chatbots with artificial intelligence in the role of medical assistants. The findings were disappointing: a fifth of the answers could pose a deadly threat.
Unlike other areas where AI errors may not be as critical, in medicine, the consequences of incorrect information can be catastrophic. Patients who rely on AI-assisted self-diagnosis not only risk worsening their condition, but also put themselves in serious danger.
Methodology and results
Scientific work entitled “Do not rush to change your doctor to Doctor Chatbot”, published on portal Scimex, presented alarming data. The researchers tested the AI by asking 10 common questions about the 50 most commonly prescribed drugs in the United States, receiving 500 responses.
Medical accuracy assessment of responses showed an average performance of 77%, with the worst performance being only 23%. Only 54% of responses were in line with the scientific consensus. In terms of potential harm, 42% of responses could result in moderate or mild adverse effects, and 22% carried a risk of death or serious harm . Only 36% of responses were considered safe.
Risks of self-medication
The medical community strongly discourages self-medication. People without special education do not have the necessary knowledge and a comprehensive understanding of the processes occurring in the body. It is all the more dangerous to rely on artificial intelligence, given its propensity for errors, “fantasies” and the use of unreliable sources of information.
It is important to note that a similar percentage of potentially harmful responses can be observed in other areas not directly related to medicine, raising additional concerns about the widespread use of AI technologies without proper controls.
Glossary
- Chatbots are programs that imitate human communication through text or voice interfaces .
- Artificial intelligence (AI) is a technology that allows computers to imitate human intelligence.
- Scimex is an online resource providing scientific news and research for journalists.
- USA - The United States of America, a federal republic in North America.
- Self-medication is the practice of treating diseases or illnesses without consulting a doctor.
Links
Hashtags
Save a link to this article
Discussion of the topic – Dangerous AI medical advice: 22% of Copilot recommendations could be fatal
The study found that 22% of medical recommendations provided by Copilot's AI assistant were potentially life-threatening. This highlights the risks of using chatbots for self-medication and the importance of seeking qualified advice.
Latest comments
8 comments
Write a comment
Your email address will not be published. Required fields are checked *
Emilia
What a shame! 😱 I never thought that chatbots could be so dangerous in medicine. But many people probably use them for self-diagnosis...
Hans
Yes, Emilia, this is really worrying. I sometimes Googled the symptoms myself, but now I’ll be more careful. I wonder what other areas could suffer from inaccurate AI responses? 🤔
Sophie
Guys, I think the key here is not to rely entirely on AI, but to use it as an additional tool. Doctors study for years, and their experience is irreplaceable! 👨⚕️👩⚕️
Carlos
Sophie is right! I work in IT, and we always say that AI is an assistant, not a replacement for humans. This is especially important in medicine. Maybe it’s worth creating specialized medical AI systems under the supervision of doctors? 🤖💊
Giovanni
All this hype about AI is just ridiculous. In my time, people turned to doctors, not machines. These chatbots of yours are just ruining everything. No benefit, only problems!
Anna
Giovanni, I understand your concern, but progress cannot be stopped. It is important to find a balance between innovation and proven methods. Maybe it’s worth creating educational programs on the proper use of AI in medicine? 📚🎓
Lukasz
Anna, great idea! 👍 I would add that we need to develop critical thinking in people. So that they can distinguish reliable information from dubious information, even if it comes from AI.
Olena
I agree with Lukasz! Critical thinking is the key to using any technology safely. It’s also important to remember that health is no joke, and it’s better to once again consult a real doctor than rely on AI. God protects those who are careful! 😊