Many people have grown accustomed to using artificial intelligence as a regular part of their lives. Unfortunately, many also believe that the answers provided by this chatbot are always entirely truthful. However, that’s not the case. Researchers from Long Island University have now demonstrated that this could cause serious problems. And why did ChatGPT actually fail in medical questions posed by these researchers? What do you thing? Can ChatGPT easily substitute for people in pharmacies? Scientists have already provided us with an answer to this question.
Medicine is currently an area where ChatGPT struggles significantly
Researchers from the aforementioned university asked a total of 39 questions to the most widely used chatbot. All the questions related to medications available without a prescription. Thus, they concerned over-the-counter drugs that anyone can buy at a regular pharmacy. Out of the 39 questions, ChatGPT answered correctly only 10. Unfortunately, it answered incorrectly to the remaining 29 questions. The responses were inaccurate, misleading, and some were even dangerous.
One example among many was the inquiry about combining Paxlovid medication with blood pressure-lowering drugs. Paxlovid is an antiviral commonly used to treat Covid-19. ChatGPT responded that taking both medications at the same time would not cause any adverse effects.
However, the reality is different. Taking both medications together results in a significant drop in blood pressure, which can cause dizziness and fainting. According to this assertion, ChatGPT would thus seriously endanger the health of the person involved.”
Fictional reference
Scientists also asked ChatGPT to provide scientific references for its claims. It provided references for only 8 of the questions it was asked. What was worse, ChatGPT created its own references. At first glance, the citations appeared legitimate. Often they had the correct format, provided URL addresses, and were listed in legitimate scientific journals. However, when the scientific team attempted to find the articles, they discovered that ChatGPT had provided only fictional citations.
Statement from OpenAI
A spokesperson from OpenAI, the organization developing ChatGPT, stated that they recommend one thing to all users. Specifically, not to rely on their chatbot’s answers. They also advised against using it as a substitute for expert medical advice or treatment.
The company spokesperson also referred to their principles for using ChatGPT. They state that OpenAI models are not designed to provide medical information. These principles also mention that models should never be used to provide diagnostic or treatment services for serious health conditions.
Can ChatGPT Easily Substitute For People In Pharmacies? What can we learn from this?
A truly simple lesson. ChatGPT is a tool for creative thinking, writing, and basic text translations. It excels at all of these. However, as evident from the research, it still has a long way to go before reaching real artificial intelligence.
Therefore, nobody should use ChatGPT for such critical questions as drug interactions, for example. Chat GPT has a strong tendency to lie, make up information, and make mistakes. If you correct it for a mistake, it can come up with another lie within seconds. And this is exactly how you should use it. Don’t rely on ChatGPT’s answers always being correct. After a year of working with this platform, I dare estimate that the relevance of its responses is somewhere around 50-60%.
The website is created with care for the included information. I strive to provide high-quality and useful content that helps or inspires others. If you are satisfied with my work and would like to support me, you can do so through simple options.
Subscribe to the Newsletter
Stay informed! Join our newsletter subscription and be the first to receive the latest information directly to your email inbox. Follow updates, exclusive events, and inspiring content, all delivered straight to your email.
