Why use AI?
AI is everywhere. And yet, it’s not always obvious. But sometimes it’s clearer that it’s being used; that reference picture you generated has an extra hand, or is missing an eye, or just feels unnatural. But artificial intelligence is getting better at its job. However, it’s not all false information and creepy- looking people, it can be extremely useful to certain professions looking to enhance their resources.
For example: Clinical Key AI (CKAI) belongs to the Elsevier company, which is the oldest and largest publisher of medical text in the world, and uses AI to retrieve information from their own books & journals, enabling the user to be able to gather information that is case-to-case specific. But why might someone turn to another well-known sites, such as ChatGPT?
The root of the problem: the internet itself
Although well-known AI search engines like ChatGPT are useful for getting quick answers, generating a picture or text to aid you for work, or summarising information, they are not always reliable, and can be confusing, with errors throughout them. This means false AI generated information is quickly spread, because there was nothing stopping the root of the problem; it is coming from the entire internet.
For example, if you asked ChatGPT to give you a recipe, it would look at all of the baking and cooking recipes published on the internet and give you a combination of all the commonalities among them. This would most likely lead to you having a recipe that is unconventional with steps that don’t follow through, and you probably won’t end up with what you wanted to achieve in the first place.
This highlights the importance of keeping control of where the information is coming from, making sure it is a credible source, like CKAI, which guardrails the information provided by limiting the data provided, and only using what is in their textbooks and journals. Also, if there is no available answer present in the book, then it will tell you that, instead of simply finding or making up an answer, meaning the spread of misinformation is potentially stopped.
Where is the information actually coming from?
One of the reasons many people don’t trust AI search engines is because they often don’t credit sources, and the information can be factually incorrect. This leads to readers being unsure as to whether the information is right or can be trusted.
However, websites like CKAI don’t use the internet to gather information; instead, they retrieve data from company books where the information is checked and controlled. This ensures the medical advice is fact-checked and doesn’t contain any influence from external sites, which may not be as easily trustable.
Potential harms
If you asked ChatGPT a specific question relating to a certain topic, with extra factors added in, would it give you a specific answer back? Most likely not.
This is because there is only so much an engine can handle. They are designed to give back information from the internet, but when it can’t find what you are looking for, it will generate an answer based on assumptions, not facts. This leads to the user getting answers that are false.
But how can this be dangerous? Well, if someone had asked for medical advice from a website like this, then they could receive false or potentially harmful ideas, presented as facts.
With CKAI, you can add in factors to ask the tool, making it case-specific, for example, instead of asking “Why would someone have a problem with this particular medicine?”, you could instead ask, “Why would a 20-year-old Caucasian female with no previous medical problems experience symptoms such as nausea and diarrhoea taking antibiotics like amoxicillin?” This gives much more nuanced and specific results, that are useful to the doctor reading up on their patients’ history.
Interview with a Chief Medical Officer
Q: What are the advantages and disadvantages of using ChatGPT compared with CKAI, and how does this affect the user experience overall?
A: The advantage of ChatGPT is that it is freely accessible, this is where the advantages stop as free and easy access does not always equate to credible information, especially when it comes to clinical workflows, where doctors and nurses and other clinicians rely on credible information to deliver care to their patients. Any non-credible information in such workflows cannot only delay care but also cause harm.
Q: Which would be more reliable? Follow up question- Is there any case in which someone would want to use ChatGPT instead?
A: ChatGPT is very useful for non-clinical workflow for example, when you want to summarize a large piece of text or to create images sometimes
Q: What are the environmental impacts of using each of them, when compared?
A: Any large language model, including AI technology consumes a lot of energy, and we need to be cognizant of equitable and sustainable use of these technologies.
Q: How does CKAI respect privacy and confidentiality compared to ChatGPT?
A: CKAI has built in privacy and confidentiality controls and is HIPAA compliant (a landmark U.S. law from 1996 that sets standards for protecting sensitive patient health information), ensuring privacy, security, and data integrity, on the other hand ChatGPT does not have the same level of controls and regulation, and can monetize private data or reuse it.
Conclusion:
Overall, AI is increasingly present, often subtly, and while it can produce errors or unnatural results, it is improving and offers valuable resources for professions like medicine. Clinical Key AI (CKAI), owned by Elsevier, retrieves case-specific information from fact-checked company books and journals, ensuring reliable medical advice. In contrast, AI search engines like ChatGPT gather data from the entire internet, which can lead to misinformation and lack of source transparency. CKAI limits information to controlled sources and provides nuanced, case-specific answers or acknowledges when no relevant data exists, helping prevent the spread of false or harmful information.
Written by Anshika