Can chatbots be trusted? what do you know.

ChatGPT’s amazing ability to answer people’s questions in instantly generated, easy-to-understand answers could be a game-changer for people seeking medical advice, but a new medical journal article says it’s not quite that simple. So does ChatGPT.

Microsoft Corporate Vice President Peter Lee has spent the past six months experimenting with the latest version of ChatGPT, an artificial intelligence platform designed to answer users’ questions with a conversational tone.

Like other so-called large language paradigms, ChatGPT, created by San Francisco startup OpenAI, is trained on huge amounts of data available on the Internet, in books, and other sources. It uses algorithms to generate words in a sequence based on the frequency with which they appear together across the Internet. Microsoft’s Bing search engine now uses ChatGPT.

Lee, a computer scientist who runs Microsoft research around the world, is particularly interested in the medical and scientific potential of ChatGPT.

previously: ChatGPT prepares to provide medical information. For better and worse.

he An article was published Thursday in the New England Journal of Medicine About the potential uses of ChatGPT for recording medical notes, offering diagnostic capabilities, educating clinicians and patients and analyzing and summarizing data. Particularly notable, he said, is the interface’s ability to provide human-like empathic responses.

We asked Lee, who is also a co-author of the upcoming book “The AI Revolution in Medicine: GPT-4 and Beyond,” about his excitement and concerns about the latest version of ChatGPT, now ChatGPT-4, which was recently released. We also asked the same questions on ChatGPT, but since its availability is still limited, the responses come from ChatGPT-3.

Question: Does ChatGPT provide reliable medical information?

for me: Personally, I don’t trust models at this level of sophistication for my personal health advice or information and would be concerned if doctors and nurses were using them to make medical decisions.

ChatGPT: ChatGPT is a language model that is trained on huge amounts of text data, including medical information. However, it is important to note that ChatGPT is not a medical professional, and its responses should not be considered as medical advice.

While ChatGPT may provide information on medical topics, it is essential to verify that information with a qualified medical professional before making any decisions regarding your health. There is also a risk that ChatGPT may provide inaccurate or outdated medical information, which could be harmful.

Question: What types of errors does ChatGPT make and why?

for me: GPT forms are generally very bad at retrieving information. It’s almost by design because GPT models don’t want to recompile their training data. If you’re asking him for a specific source, like a reference to a paper, he won’t be able to bring the exact source to life, so he makes something up, he’s “hallucinating”.

ChatGPT: ChatGPT may make errors such as providing inappropriate or incorrect responses, repeating itself, or producing responses that do not match the context of the conversation.

These errors can occur because ChatGPT is trained on huge amounts of text data, including unverified and potentially biased information, which can lead to incorrect or outdated information.

In addition, ChatGPT responses may be affected by the quality of the input it receives, which may result in irrelevant or off-topic responses.

Question: How is ChatGPT-4 better than ChatGPT-3?

for me: We reach a different level of intelligence. The main thing is you have this AI model that has a lot of cognitive capabilities and you have a chat interface. It ends up being really natural to interact with.

ChatGPT: It is reasonable to expect improvements in areas such as scale, accuracy, multilingual capabilities, understanding of context, and consistency.

Question: ChatGPT sometimes looks like a real person. Does ChatGPT have a “theory of mind”, which is the ability to imagine other people’s mental states?

for me: You can’t do this. It’s a machine. But you can ask GPT-4 what the patient (in a given situation) would describe as thinking and feeling. He is able to assess that state of mind. … It translates into a feeling of empathy for patients.

ChatGPT: No, ChatGPT does not have a theory of mind.

ChatGPT is a machine learning model that uses statistical patterns and algorithms to generate responses to natural language inputs. It has no consciousness, emotions, or beliefs, nor does it have the ability to understand or infer the mental states of human users.

While ChatGPT is able to generate responses that may be empathetic, engaging, and conversational, it does not possess a theory of mind, which is a critical aspect of human cognition and social interaction.

Question: Should ChatGPT be used in medicine?

for me: Society has a choice. We can try to slow down or turn off this technology, because we don’t fully understand the risks. The second argument is more fatalistic, saying, “He’s unstoppable. We might as well try it.” The third option, which I think is the only rational one, but it should be made intentionally, is to think carefully about how we can integrate this technology into our lives so that we can achieve things that neither humans alone nor AI alone can do. . I hope the medical community and the public will choose the third option.

ChatGPT: ChatGPT should not be relied upon solely for medical advice or diagnosis. While ChatGPT may be able to provide information on medical topics, it is not a medical professional, and its responses should be considered informational only.

However, ChatGPT can be a useful tool for gaining a general understanding of medical concepts and terminology. In any case, it is essential to verify any information provided by ChatGPT with a qualified medical professional to ensure that it is accurate and applicable to your situation.

Contact Karen Weintraub at [email protected].

USA TODAY’s health and patient safety coverage is made possible in part by a grant from the Masimo Foundation for Ethics, Innovation, and Competition in Health Care. The Masimo Foundation does not provide editorial input.