Read the following text and answer the quiz below.
[1] Next time you’re considering consulting Dr ChatGPT, perhaps think again. Artificial intelligence chatbots might be able to ace most medical exams. But they do not give humans better health advice than what people can find using traditional methods, according to a study published last month.
[2] “Despite all the hype, AI just isn’t ready to take on the role of the physician,” Rebecca Payne said. She is from the University of Oxford and is a co-author of the study. “Patients need to be aware that asking a large language model about their symptoms can be dangerous. She added that chatbots can give incorrect diagnoses and fail to recognise when urgent help is needed.
[3] The study was published in the Nature Medicine journal. The British-led team wanted to learn how successful humans were when they used chatbots to identify their health problems. Their study also investigated whether chatbots could pinpoint when a human needed to see a doctor or go to the hospital.
[4] The team presented nearly 1,300 UK-based participants with 10 different scenarios. These included a headache after a night out drinking alcohol or a new mother feeling exhausted. Then the researchers randomly assigned the participants one of three chatbots: OpenAI’s GPT-4o, Meta’s Llama 3 or Command R+. There was also a control group that used internet search engines.
[5] People using the AI chatbots were only able to identify their health problem around a third of the time. Around 45 per cent figured out the right course of action to take. These numbers were no better than those of the control group.
[6] The team noted a gap between the disappointing results and AI chatbots’ extremely high scores on medical exams. They blamed it on a communication breakdown, saying that real humans often do not give chatbots all the relevant information. They may also struggle to interpret the options a chatbot presents, misunderstand its advice or simply ignore it.
[7] One in six US adults asks AI chatbots for health information at least once a month, researchers said. That number is expected to increase. “This is a very important study as it highlights the real medical risks posed to the public by chatbots,” David Shaw said. He is a bioethicist at Maastricht University in the Netherlands who was not involved in the research.
Source: Agence France-Presse, February 9
Content provided by British Council
