Using AI—Who is the Expert?

I’m in the process of completing an excellent review on AI, AI for Work and Life, an eight-part webinar series put on by my alma mater, the University of North Florida. Each session is at least an hour long, and is led by one or more subject matter experts in different aspects of the use and, yes, misuse of this world-changing technology. We’re getting the latest and best understanding of who can or should use AI, what some of the many uses of it can be, and how to use it accurately, responsibly, and ethically.

I couldn’t possibly cover all of the terrific, thought-provoking content I’ve taken in thus far We’ve considered both the benefits and shortcomings of AI usage. As a content creator myself, one thing really stood out. A huge problem is that sometimes the AI engine will make up content that it thinks the user wants.

For example, as noted in a recent NY Times article, there have been over 500 instances of attorneys being cited by different legal systems for presenting legal briefs with citations provided by ChatGPT or some other AI engine that were found to be nonexistent—hallucinations in AI parlance. Perhaps the lawyer involved was laboring under a deadline. Perhaps they were just lazy. Either way, it points out a big problem with using AI—hallucinations. But the good news is that those fabrications can be avoided when in the hands of someone who knows the subject matter at hand and will take the time to both check resources and create AI queries that minimize false output.

The other interesting thing brought up was the fact that people in many areas of business other than the legal profession are now attempting to create content with AI without knowing its downsides and how to minimize errors and the production of unfocused, boring content. After realizing the mess they’ve made, they then bring that mess to a subject matter expert to clean up. That takes a lot of work to do! They would’ve been better off to have just brought their need or problem to the expert in the first place. That would’ve saved them a lot of time and money.

At present I use AI mostly for research purposes. I’ve already run into some hallucinated answers—research papers that didn’t exist, content that was supposed to be in a certain magazine article that wasn’t. I knew from the start that I am ultimately responsible for the accuracy of any content that I produce, But now I’m learning how to design my queries in such a way that those made-up responses are minimized.

The focus in my work remains on providing healthcare and wellness content that is both engaging and accurate. AI can assist me greatly in that area, but someone has to be the subject matter expert. That’s something that AI cannot do or be. As a proud RN with many years of healthcare clinical and business experience, that someone is me. I’d rather write nothing at all than produce something that was somehow inaccurate or intentionally misleading. That’s my pledge and my promise.