A cramp in the stomach, a persistent cough or a strange spot on the toenail: it is not new that people google on Google when they have complaints. You would think that AI chatbots now promise even more possibilities for self-diagnosis. But there are reservations.
With the increasing popularity of AI chatbots like ChatGPT, the possibilities for digital self-diagnosis seem to have increased, you would think. Research shows that the medical competence of artificial intelligence is still viewed with great reservation.
Scientists from Würzburg investigated people’s reactions to AI-generated medical advice. “We were not interested in the technical competence of the AI, but rather in the question of how the AI output is perceived,” says Moritz Reis from Julius Maximilians University about the study published in the journal ‘Nature Medicine’.
Study with over 2000 subjects
To do this, the research team divided over 2,000 subjects into three groups that were given identical medical advice. The first group was told that the recommendations came from a doctor. The second group credited an AI-powered chatbot as the creator, and the third group assumed that the advice came from a chatbot but was vetted by a doctor.
Subjects rated the recommendations for reliability, understandability, and empathy. Once they suspected that an AI was involved, they perceived the advice as less empathetic and trustworthy. This was also true for the group that believed a doctor had rated the AI recommendations. Accordingly, they were less willing to follow these recommendations. “The effect of anti-AI bias is not huge, but it is statistically significant,” Reis said.
The cognitive psychologist partly explains the AI skepticism with stereotypes: “A lot of people believe that a machine cannot be empathetic.” But when it came to understandability, all three groups rated the advice the same.
AI skepticism is highly relevant for medical professionals
The observed AI skepticism is important for the research group, as AI is playing an increasingly important role in medicine. Numerous studies are currently being published on new AI applications. This makes public acceptance all the more important, says Reis: “The question of the future use of AI in medicine is not only about what is technically possible, but also about how far patients will go.” necessary. “In addition, other studies have shown how important it is for patient trust that the human doctor always has the final decision-making authority together with the patient,” Reis emphasizes.
The scientist finds transparency particularly relevant: “This means, for example, that an AI not only makes a diagnosis, but also explains in an understandable way what information led to this result.”
Chatbot diagnostic hit rates vary
The quality of these results has been scientifically investigated for a long time, with varying degrees of success. In 2023, for example, ChatGPT demonstrated a high degree of diagnostic accuracy in the “Journal of Medical Internet Research”: tested with 36 case studies, the chatbot made the correct final diagnosis in almost 77 percent of cases. According to a Dutch study, the diagnostic skills in the emergency room even approached those of doctors. Armed with anonymized data from 30 patients treated in a Dutch emergency room, ChatGPT made the correct diagnosis in 97 percent of cases (Annals of Emergency Medicine, 2023).
In contrast, a study published in 2023 in the journal Jama found that the chatbot correctly diagnosed only 27 cases out of 70 medical case studies. That’s just 39 percent. A study presented in the journal Jama Pediatrics found that this hit rate is even worse for diseases that primarily affect children.
A recent study published in the journal “Plos One” has now investigated whether ChatGPT could be useful in medical education. Ultimately, the chatbot not only draws on a vast knowledge base, but is also able to convey this knowledge in an interactive and understandable way, according to the research team from the Canadian London Health Sciences Centre.
Opportunities in medical education
The group fed ChatGPT 150 so-called case challenges from a database of medical case histories describing symptoms and disease progression. Both prospective and current physicians are asked to make a diagnosis and develop a treatment plan using an answer choice procedure.
In this test, ChatGPT was correct in just under half of the cases (74 out of 150). The study found that ChatGPT has difficulty interpreting laboratory values and imaging tests and misses important information. Accordingly, the authors conclude that ChatGPT in its current form is not accurate as a diagnostic tool and that caution should be exercised when using the chatbot as both a diagnostic tool and a learning tool.
“The combination of high relevance and relatively low accuracy argues against relying on ChatGPT for medical advice, as it may present important information that could be misleading,” the study says – a warning that most likely applies to ChatGPT as well. This is true for medical laypeople who use the chatbot for digital self-diagnosis.
ChatGPT itself emphasizes that it is not suitable for this. When asked about its diagnostic qualifications, the bot answers: “I am not a doctor and have no medical training. I can provide information on medical topics, give general advice, and answer questions, but I cannot make medical diagnoses or provide professional medical advice. If you have any health concerns or questions, you should always consult a doctor or qualified healthcare provider.”
Source: Krone
I am an experienced and passionate journalist with a strong track record in news website reporting. I specialize in technology coverage, breaking stories on the latest developments and trends from around the world. Working for Today Times Live has given me the opportunity to write thought-provoking pieces that have caught the attention of many readers.
