It seems so easy–just plug a question into an AI chatbot like ChatGPT and, in no time at all, get a complete answer you can drop into a blog post or article. With content writing reduced to a few minutes, you can then move onto other tasks.
But what if your chatbot is wrong?
Accuracy in AI Chatbots
We’ve all heard wild stories about the inaccuracies of AI applications, some of which may stem from earlier incarnations. In reality, AI chatbots like ChatGPT have improved since their early days; recent estimates say that the latest version of ChatGPT, in particular, gets things right about 89% of the time. (Being correct is one thing; the tone of the content is another. That AI content always seems to sound like a machine wrote it is a topic for another day.)
In a school assignment, 89% is pretty decent. But if you are sharing information with your customer or client base, it’s not so great.
What are the chances your content lands in the 11% that’s incorrect? Are you willing to take that risk?
There’s also the issue of so-called hallucinations, defined by Chatbase as “instances where the model generates factually incorrect or nonsensical information with a confident tone.” The latest version of ChatGPT hallucinates 28.6% of the time and “often fails to indicate uncertainty.” In other words, even if its answer sounds amazing and insightful, there’s a nearly 30% chance it’s completely false.
Concerns about chatbots are especially acute when it comes to what is colloquially known as Your Money or Your Life topics. (YMYL) This is the kind of content that dispenses advice in areas like personal finance, health, or nutrition, where accuracy is absolutely essential. Incorrect information here can affect a person’s wealth and health, hence the YMYL label.
Beyond YMYL topics, incorrect information is bad for any business because it erodes trust, and no business can afford that.
Even companies in the AI business know that chatbots don’t have a great record in content creation. The article I cited above from Chatbase (which is, itself, an AI tool) notes that users should always double-check information provided by chatbots. It also advises users to remember that chatbots are tools for brainstorming or generating ideas, not “definitive source[s] of factual information.”
That statement confirmed any biases I had about AI. These so-called large language models (LLM) scrape the internet for content–generated by humans–to use in their replies, without consent or attribution. Because of this theft, I have never been a fan, but knowing the “content” AI chatbots produce is of dubious value reinforces my view that they must be used with caution for copywriting.
Of course I can’t base my opinion on my own preconceived notions. I know that to make the case for human fact-checking of AI content, I have to gain some understanding of what chatbots do and how well they do it. I had never used a chatbot until writing this blog post, but decided to try one and assess the results for myself.
Querying ChatGPT
For the purposes of researching this article, I chose ChatGPT since it is the chatbot I have heard most about. I opted for a topic on the lighter end of the YMYL spectrum, that is, not life or death but not without serious implications.
I had been reading recently about the potential health benefits of omega-3 fatty acids. One article I stumbled upon mentioned eye health so I decided to go down that avenue with ChatGPT to see what it would say. I asked the following question: “Do omega-3 fatty acids help with eye health?” Here was its response:
Point 2 was of interest to me since I have had issues with dry eye.
While it does not say that omega-3 fatty acids will cure dry eye syndrome, it does say definitively that they improve tear production and reduce inflammation. And compare point 2 to point 3–the answer to the latter “suggests” a benefit, where point 2 declares it. Finally, there is the note on sources and recommendations for supplements and foods. Put a pin on this information because it will be relevant later.
So we’ve seen what the machine says, but what does the research say?
The Research Record
I provide detailed excerpts and citations for the research below, but here are the highlights. (For the record, I used academic journals as my resources. While I have cited just three articles, their research is solid and shows that ChatGPT is, in this case, hallucinating and should not be making such a definitive claim about omega-3s being a dry eye treatment. The sources from 2022-2025, so they are recent.)
- There is nothing conclusive yet. A few articles noted that there is insufficient evidence for omega-3 fatty acids as a dry eye treatment and that more research is needed.
- Supplements may not be effective. Some studies noted that topical use of omega-3 oils might be far more effective than oral use. (Don’t do this without the supervision of an optometrist or ophthalmologist!) The distinction between topical and oral is important because of the next point.
- People tend to self-medicate. There has been a lot of positive press about omega-3 oils for a variety of ailments. According to one article referenced below, people tend to self-medicate. That is, if they hear that these oils might help with dry eyes, they’ll just start taking them, which is easy enough to do since omega-3 supplements are widely available and do not require a prescription.
- Past research has a high level of bias. An analysis of previous studies into omega-3 fatty acids and dry eye syndrome shows they all had high or unclear levels of bias. Why? Because pharmaceutical companies stand to make money from omega-3 supplements; touting them as a dry eye treatment helps their bottom line.
So ChatGPT is telling us that “Omega-3s improve tear production and reduce inflammation, helping with dry eyes and irritation.” Yet the research shows: there is no clear proof that tear quality or inflammation changes with omega-3s; any research that makes such a claim is likely biased; and oral use of omega-3s might be far less effective than topical use.
You might ask, “What’s the big deal? An omega-3 supplement is unlikely to hurt anyone and might benefit them in other ways.”
While it’s true that the claims I’ve outlined above are not on par with something like “hydroxychloroquine cures Covid,” there are potential harms:
- Omega-3 supplements are expensive and while they have many potential health benefits, dry eye might not be among them. So anyone spending money on these supplements as a dry eye treatment might find they’ve wasted their money.
- Or, knowing the wider range of health benefits of omega-3 oils, someone might forgo other dry eye treatments, figuring omega-3s will help them with lots of things, including their dry eyes.
Bottom line: people might spend money on a supplement that does not treat the problem and, simultaneously, stop a treatment that is more beneficial because they believe the as-yet-unproven guidance that omega-3 oils will help.
As the saying goes: your money or your life.
You probably don’t want your business associated with suspect advice like this. That is why you need AI fact checking. Chatbots are not infallible and the mistakes they make will not only affect the people who heed their advice, but the businesses that share it.
For further reading on the reliability of AI chatbots, see Can You Trust AI Chatbots? I found different answers from ChatGPT for very similar questions; one of the omissions was particularly shocking.
Photo of desk by Nicola Nuttall on Unsplash.
The Full Research
Here are some excerpts from the research mentioned above. Emphasis, in the form of italics, is mine.
More Research Is Needed
As noted in the Journal of Education, Health and Sport in 2024:
“Omega-3 fatty acids may have benefits in the treatment of dry eye syndrome…However, the literature does not agree on the efficacy of their use in the treatment of dry eye syndrome, and thus there is insufficient evidence to bring omega-3 acids into the standard of treatment for dry eye syndrome…There is a need for further long-term studies to standardize conclusions about the efficacy of these acids in reducing dry eye symptoms.”
And, from the Journal of Young Pharmacists in 2024:“Omega-3s may alleviate symptoms of dry eye syndrome by reducing inflammation in the eye’s tear glands and improving tear quality.”
There is a maybe here, not a clear statement.
Topical Use vs. Oral Use
As noted above, it seems the real benefits of omega-3 fatty acids in eye health might be delivered topically rather than orally. The Journal of Young Pharmacists, cited above, suggests that topical use of omega-3 oils might be more effective than oral use. (Again: do not try this without the supervision of an optometrist.)
Topical use is also mentioned in a 2023 article from Acta Ophthalmologica: “It must be pointed out that the two most recent trials from 2020 explored the use of topical omega-3 rather than oral supplementation (Downie et al., Laihia et al.)… All of these results are promising that the future of omega-3 fatty acid supplementation may, in fact, lie in topical use as opposed to oral intake.”
That is, the jury is very much out on the benefits of oral supplements, but people may use them anyway because of something they read on the internet.
The article also notes the tendency for people to self-diagnose in this area and start omega-3 supplements independent of advice from an eye care professional.
Bias in Research
The Acta Ophthalmologica article mentioned above is a meta-analysis, that is, an article that looks at the results of previous studies. As such, the authors considered the degree of bias in previous studies: ““The level of bias in the included studies was generally unclear or high, and there were no trials without any high or unclear risk of bias.”
It goes further to note one of the reasons for the bias, which comes down to money: [T]here is a significant economic burden associated with dry eye disease and correspondingly, there is considerable motivation for pharmaceutical companies to promote their treatments. In fact, a report from 2020 has estimated that the dry eye disease market is predicted to reach USD 6960 million by 2026.”
Its final conclusion also notes that: “there is insufficient evidence to advise or refute the use of omega-3 fatty acids in the treatment of dry eye disease.”
Sources
Ahmad, Sarfaraz, et al. “Incredible Use of Omega-3 Fatty Acids: A Review on Current Use and Future Prospective.” Journal of Young Pharmacists. Online. 2 June 2024. Vol 16, No. 2, pp. 177–86. Crossref, https://doi.org/10.5530/jyp.2024.16.24.
Chatbase. “Is ChatGPT Accurate? Latest Data & Reliability Tests (2025).” Chatbase. https://www.chatbase.co/blog/is-chatgpt-accurate, Jan.21, 2025.
Martyka, Anna et al. “Exploring the Efficacy of Omega-3 Fatty Acid Supplementation in Dry Eye Disease: A Comprehensive Review.” Journal of Education, Health and Sport. Online. 18 January 2024. Vol. 53, pp. 163-173. [PDF link available to full article.]
O’Byrne, Ciara and Michael O’Keeffe. “Omega-3 fatty acids in the management of dry eye disease—An updated systematic review and meta-analysis.” Acta Ophthalmologica. Online. 22 September 2022. Vol. 101, Issue 2. pp. e118-e134.