M. Sitki Copur, MD
Morrison Cancer Center
Since Artificial Intelligence exploded into our daily lives, millions of people have turned to the internet and found information designed to simulate conversation with human users for health advice. Patients and families sometimes rely on chatbots to obtain information about their health conditions, and the market for these chatbots has grown rapidly — ChatGPT, Google's Gemini, Microsoft's Co-Pilot, MetaAI, Claude, Grok and Perplexity. All have strengths and weaknesses. They seem to have the potential for revolutionizing cancer patients’ access to information.
However, there are many times when artificial intelligence just doesn’t quite get everything right. This is especially true when you want information about your specific type of cancer. That’s why your local healthcare providers are a better source for answering your questions and helping you get the best possible information.
Artificial intelligence works by generating answers to some of the most Googled facts about cancer, such as, “What are the most common cancer types?” “What exactly is cancer?” However, what makes artificial intelligence models a disruptive innovation is not their ability to answer simple questions, but their lack of ability to formulate responses and advice to complex questions and conversations.
When it comes to healthcare information, artificial intelligence does not always get it right. Many challenges exist in accuracy, reference quality and readability of health information.
“Hallucination” in artificial intelligence refers to a situation in which a chatbot model generates false or misleading information, presenting it as accurate. Hallucinations are often caused by factors such as insufficient training data, incorrect assumptions or biases within the training data. All artificial intelligence models are prone to hallucinations.
When you are using artificial intelligence, be aware that:
- The information does not credit sources.
- There is no single answer to a question (e.g., the answer differs when using same prompt, as well as the preceding conversation and/or context.
- There might be some incorrect statements that sound real. Many users can see the information from .com sites, but unlike government (.gov) or academic (.edu) sources, these .com websites are not held to stringent standards for medical accuracy. So, they are not generally as authentic and are often considered unreliable. They may prioritize commercial interests over accuracy and scientific evidence.
With the rapid advancement and extensive interest in artificial intelligence chatbots, patients and caregivers are increasingly relying on artificial intelligence for medical advice. Yet there is an urgent need to improve the quality of information in these artificial intelligence chatbots to ensure safe and equitable access to healthcare knowledge. This will require regulators and healthcare professionals to be involved in developing standards for minimum quality and to raise patients’ awareness of current limitations.
If you or a loved one is given a cancer diagnosis, please do not panic or Google for information. Ideally, it is best to reach out to your healthcare provider or oncologist whenever possible. In addition, you can call the Morrison Cancer Center at 402-460-5899.
