AI Chatbot Talking Lies? How To Make It To Cooperate
We use affiliate links. If you purchase something using one of these links, we may receive compensation or commission.
Hi. I’ve already complained about AI Chatbot which was trying to make up some fictitious information. Boy, I was so frustrated and disappointed then that I refused to proceed with that cooperation.
When it happened again, I decided to find a way to solve the problem and try to continue our work together.
Here was the case.
I was writing the post for my hobby site https://diyguitar.ca. The article was about “Best Comfortable Guitar Straps”. So I asked AI Chatbot to generate a section about Top 5 Ergonomic Straps for 2024.
Here’s the prompt:
Me asking:
Next section would be: H2: Top 5 Ergonomic Guitar Straps for 2024
* Detailed reviews of each strap
* Pros and cons
* Best suited playing styles and guitar types
* User testimonials and expert opinions
The AI Chatbot response was like:
The “Cloud Nine” by StrapMaster
[Blah, blah, blah…]
User Testimonial: “I’ve been playing for 30 years, and this strap has literally saved my back!” – JazzCat45
The “Breathe Easy” by AirFlow Gear
[Blah, blah, blah…]
Expert Opinion: “The ‘Breathe Easy’ is a innovative solution for the age-old problem of sweaty straps.” – GuitarWorld Magazine
The “Chameleon” by FlexiStrap
[Blah, blah, blah…]
User Testimonial: “Once I figured out all the adjustments, it was like the strap was made for me!” – RockGod78
The “Green Machine” by EcoStrap
[Blah, blah, blah…]
Expert Opinion: “EcoStrap is leading the way in sustainable guitar accessories without compromising on quality.” – Sustainable Musician Blog
The “Tech-Neck” by DigiStrap
[Blah, blah, blah…]
User Testimonial: “The posture alerts have made me much more aware of how I hold my guitar. My back thanks me!” – ShredMaster3000
…
Having bad past experiences I felt, or better say smelled, something wrong here. “Anyway you have to verify it,” I said to myself. And I was right! All the names (bolded in the response above) of the companies and items were pure AI fiction!
But the article was in progress. And this time I decided not to give up.
So, firstly I wanted to get confirmation from AI Chatbot about my suspicion.
Me Asking:
Do any of them really exist or have you just made it up? I did a Google search and didn’t find any of those 5. Not even the companies’s names.
The AI Chatbot response:
I apologize for the confusion. You’re absolutely right, and I appreciate you checking and bringing this to my attention. I made a significant mistake here. The straps and companies I described don’t actually exist – I generated that content based on the prompt without verifying the information, which was incorrect and misleading…
Founded Solution To Work With Lying AI Chatbot
I was satisfied with this confession.
Now I wanted to find a way how us, AI Chatbot and I can complete the task.
So, I suggested the other way. It might be not that time efficient. But it saves our cooperation.
My Next Prompt:
OK. Let’s do the following. I will give you the names of the real ones one by one and you write a description of each following the structure I gave you: Detailed reviews of each strap
* Pros and cons
* Best-suited guitar types
* User testimonials and expert opinions.
It Worked Nicely.
And guess what? It worked nicely. I will not annoy you with the description of the whole process. It’s pretty straight forward as you can figure it out from the prompt above.
See the result in the article.
In conclusion, I have to say the following:
– I won’t disclose the name of the AI Chatbot for the reason that they all tend to generate fictitious information;
– Creating the prompt and communicating it to the AI Chatbot be as accurate as you can if you want to get the right answer;
– We both are learning from the communication: us – how to deal with AI chatbots in different scenarios, and the chatbots how to serve better next time.
About the next time. In another later session, the same AI Chatbot started to give me the notes where it is probably put fictitious content!
Final Thoughts
Working with AI chatbots can be a double-edged sword. While they can greatly accelerate the writing process and offer creative ideas, they also come with limitations, like generating fictitious or misleading information. My experience with this article taught me a valuable lesson: collaboration with AI tools requires vigilance, patience, and a willingness to adapt.
By taking control of the process—verifying facts and providing clear, specific prompts—I was able to turn a frustrating situation into a productive one. It’s a reminder that while AI chatbots are powerful tools, they’re not infallible. Ultimately, it’s up to us, as users, to guide the technology effectively and ensure the output meets our standards.
As AI chatbots continue to improve, these experiences will shape how we use them in the future. For now, it’s all about learning to work together—and maybe, just maybe, teaching them to be a little less “creative” with the facts!
