Have you reached the point of frustration with AI chatbot recommendations? Me too.
We all know that ai chatbots don’t always give us 100% accurate answers
But that’s not the biggest problem; the real issue is that ChatGPT and Google AI recommendations are sometimes intentionally misleading us.
Let me give you an example.
This is me, on aregular Tuesday. Scrolling the internet and trying to figure out which Korean sunscreen is best for my skin type. So I did what every millenial in my position would do: I asked Google AI.
Why? It’s fast. And convenient. It saves the time I’d otherwise spend scrolling through a number of online articles and listicles. Reading comparisons, getting into technical details and chemical ingredients. That’s not always fun. And I’d rather skip checking on individual customer reviews – unless it’s really important.
In just a few seconds the Google AI generated a lot of sunscreen choices.
But there’s one problem with these AI chatbot recommendations.
They’re often pulled from PR-driven content or e-commerce sites directly. As you can see in the list of sources the top recommendations are cited from websites selling Korean products like Skinorea.com.
The product recommendations came from this blog post “Best Korean Skincare Products in 2026” – doesn’t even have an author – it’s just a random listing on an e-commerce storefront.
And people still talk about AI using quality data sources?
Or that AI is better at doing research than humans?
Sounds like a joke to me …
In other words, you might be getting influenced by marketers without even realizing it.
And I know this because I’m one of them. I am a marketer, not a robot, just to be clear.
As a content writer, I get paid to write blog posts that highlight a brand’s unique benefits, compare products, and make things sound more appealing to influence the reader’s decision-making. It’s not likely a brand will pay for content that talks negatively about them. So, the flaws are usually skipped or hidden in an article or review.
That’s the content writer’s job nowadays. And, it’s gone a step further. We’re no longer just writing for people; we’re writing content specifically to get cited by AI.
So yes, most content published today is increasingly optimized for tools like ChatGPT, Gemini, and Perplexity AI. Writers no longer focus only on SEO, but the goal is to dominate these AI citations.
Why? Because people are spending more time asking chatbots for solutions, recommendations and tips than scrolling on social media. In a way, chatbots are becoming the new influencers.
And people trust them, because the assumption is:
“AI chatbots scan the internet, analyze complex data, read every statistic and article on this topic, and give me the best answers.”
WRONG. That’s not really how it works.
Chatbots can only summarize the online content that already exists online. Sometimes, it’s not even up to date. Other times, it’s pure PR content that gets picked up. What exists online on company websites is often shaped by what the brands, marketers, influencers, and magazines publish. And a LOT of what you read is sponsored content (even though it’s not always properly disclosed).
AI citations and answers can be strongly influenced by inaccurate sources
Here’s how the loop usually goes:
- Brands pay for exposure (through blogs, influencers, paid reviews, PR placements)
- Positive content gets published (and sometimes competitors get subtly criticized)
- AI models pick up on the overall sentiment
- Recommendations start reflecting that sentiment
And just like that, certain products become “the best” in the market, but not necessarily because they are reviewed by users, but because they’re the most talked about online.
You don’t even realize it… until you buy something that doesn’t live up to the hype.
So what’s the alternative?
I’m not saying you should completely ignore AI recommendations. But you need to be more skeptical of the answers it gives.
Start by asking for reputable sources. Then actually click and read them to see where the information was puled from. Does it look like a PR campaign for a brand?
If the source is something like:
“A vs B: Why Our Mattress Is Superior”, that’s clearly a biased take and you can’t trust it.
A stronger source would be something independent. A study, a comparison across multiple brands, or a reviewer who isn’t sponsored.
Another underrated approach? Go where real people meet to talk and discuss issues: forums, niche groups, and especially Reddit.
Personally, I trust Reddit threads way more when I’m researching things like hotels or travel locations. You get details that no polished blog post will ever include:
- whether the walk to the hotel involves a steep hill
- if the area feels unsafe at night
- is it located on a dodgy street
That kind of lived experience is hard to fake. And it’s definitely not something AI summaries are good at capturing.
At the end of the day, the information is out there. But we can’t be too lazy about it. Sometimes, doing your own digging and reading will still give you better results.
Curious to know: have you ever bought something based on a chatbot recommendation and regretted it?
Leave a Reply