If your doctor uses chatgpt, your cancer could go unnoticed

Chatgpt is more than a simple conversation tool, AI has become a precious help for many users. During the conference for the presentation of its 5ᵉ version, its creators revealed that AI had also accompanied many sick people in understanding their symptoms. She even managed to diagnose diseases better than real doctors.

Only, a new study breaks this good news and reveals that caregivers who use artificial intelligence would be able to be wrong. Diseases like cancer could go unnoticed.

Chatgpt

Chatgpt is the Openai chatbot, based on the GPT artificial intelligence model, making it possible to answer all kinds of questions or requests. Available in free online version.

• License:
  Free license
• Author :
  OPENAI
• Operating systems:
  Windows 10/11, Apple Silicon macOS, online service, Android, iOS iPhone / iPad
• Category :
  IA

Chatgpt would be a bad doctor

The Lancet has just unveiled a study that is cold in the back. Doctors who use artificial intelligence in their diagnosis are more likely to make mistakes, especially in detection of cancer. “”The main evaluation criterion was the variation in the detection rate of adenomas (cancer anomaly) of the standard colonoscopy not assisted by IA, before and after exposure to AI. “

The study followed doctors from four endoscopy centers in Poland, with the participation of researchers from several countries, and that’s what they discovered:
The cancer detection rate increased from 28.4 % to 22.4 % with AI, a drop of 6 %. In addition, diagnoses have also been influenced by age (60 or over), sex and use of AI.

Doctors, accustomed to delegating part of the machine observation, would risk losing in vigilance and precision, which could lead to missed diagnostics. In the long term, this raises the question of dependence on AI tools and the impact on training, skills and autonomy of health professionals.

Diagnose with Chatgpt, the false good idea?

However, as we said, AI has greatly helped users in the field of health. The solution is not to exclude Chatgpt or other chatbotsbut to use it intelligently. Always ask for your sources at AI and check its reliability. If for example she quotes a forum, leave it aside and favor scientific studies made on a large number of participants.

If the answer is based on a forum or isolated testimonies, process this data with caution. On the contrary, favor references from serious scientific studies, published in specialized journals and carried out on a large number of participants.

In addition, it is essential to never stick to what an AI says. Even if it helps you make a diagnosis or better understand a health problem, this does not replace human expertise. If you think you have identified a disease or risk thanks to a chatbot, always make an appointment with a doctor.

Want to save even more? Discover Our promo codes Selected for you.