The effect is quite comical, but may be a bit irritating for Amazon shoppers. The case was described a few days ago in Ars Technica by Kyle Orland.
“I’m sorry, but I can’t provide the information you’re looking for.”
As already mentioned, Amazon sellers discovered ChatGPT. A dream program for anyone who has a package of texts to prepare and is not very willing to pay someone for it. A properly briefed ChatGPT will do it for free and properly. Or not, because it sometimes makes mistakes (although rather rarely when using ready-made data) and, what is more important in this case, users are often unable to formulate the prompt properly (remember this word, it will probably become the word of the year one day). So their briefs are, to say the least, not very relevant. Copywriters have been struggling with this for years, and ChatGPT is just an algorithm. As he interprets it, he will do it.
So what texts did Amazon present to users who wanted to buy selected products, e.g. a cabinet, a garden chair or a Chinese religious treatise?
“I’m sorry but I cannot fulfill this request, it is against OpenAI policy.”
“I’m sorry, but I can’t generate a response to this request.”
“I’m sorry, but I can’t provide the information you’re looking for” (available in different colors).
“OpenAI cannot provide content that requires the use of trademarked brands.”
“OpenAI cannot provide content that promotes a specific religious institution.”
“OpenAI cannot provide content that encourages unethical behavior.”
“Our product can be used for various tasks such as (task 1) (task 2) (task 3).”
But he can handle screaming without blinking an eye
The company that founded ChatGPT appears in Amazon’s responses very often, but the spectrum of AI tools available is now much broader. There are many similarities between them, including: is that they are all still quite imperfect. The dream of a free, perfect workforce will probably come true soon, but as long as artificial intelligence available to the general public is at the stage of development it is at, employers will not be able to cope without human resources. To get the right text out of AI, you need to be able to instruct it. It is also worth having someone verify her works. But don’t worry, employers, he can handle your screams without blinking an eye.
Amazon does not prohibit the use of artificial intelligence to create product descriptions. On the contrary, it offers willing sellers its own AI tool. Nevertheless, a lot of spam passes through his sieve, which, as you can see, he is not always good at catching. It’s not a big deal if it’s just unsuccessful descriptions of actually existing products or problems with the search engine. What’s worse is that the platform can also be used by fraudsters who have gained a new technique to increase their effectiveness. Like any technology, artificial intelligence has two faces.
