What AI “wrote” was so outrageous that some sellers’ listings were removed from Amazon!

As the concept of AI (artificial intelligence) has become more and more popular in the past two years , all walks of life have invested in the research and use of AI.

The e-commerce industry is no exception. As an industry giant, Amazon has been continuously investing and taking actions in AI. A while ago, it launched four AI functions to help buyers find better-fitting clothes and reduce sellers' return rates.

Amazon's original intention of developing and encouraging sellers to use AI was to enable sellers to operate better and more conveniently on the platform. However, recent events seem to have backfired.

Recently, foreign media reported that some new product descriptions on Amazon caused an uproar, because some product listings showed jaw-dropping descriptions such as " I can't meet this request, it violates OpenAI's policy ", "Sorry, I can't generate a response to this request", "Sorry, I can't provide the information you are looking for".

These products have one thing in common - the listings are generated by written AI. It is unknown whether the sellers posted them intentionally or unintentionally, but these statements appear on Amazon so openly.

According to the report, the mistakes involved a wide range of categories, from lawn chairs to office furniture to religious books.  

For example, a dressing table with a small green sticker has a product description that reads " Sorry, I can't fulfill this request because it violates OpenAI's usage policy ." Some people speculate that this listing is "optimized" by the seller for search engines so that the product can be more easily "discovered" by consumers.

From the seller's perspective, this idea is understandable, but from the consumer's perspective, the appearance of such listings makes people wonder whether there is anyone who reviews Amazon's product displays. After all, as an online shopping platform with a rich variety of products, Amazon does have some product listings and reviews that are more or less "watered down".

In response, Amazon told users that they have never relaxed on "ensuring that users have the most trustworthy shopping experience"; they will delete all problematic product listings to ensure the credibility of the platform.

As Amazon assured, some of the items in this situation are no longer available, but some still exist.

Coincidentally, this is not the only embarrassing thing that happened to Amazon regarding AI.

I wonder if any of you remember the article titled "Exaggerated negative reviews! Amazon's AI function has become a "heartache" for sellers" pushed by the official account on December 21. It said that Amazon's AI review summary function exaggerated the negative information in the reviews, causing sellers a lot of headaches.

After New Year’s Day this year, some sellers said that their listings were changed by Amazon using AI during the New Year’s Day, and they tried to refresh them using templates but failed. In the end, they had to delete the entire listing and then recreate it.

In fact, not only ChatGPT and Amazon AI, but other generative AIs I use also occasionally have similar problems such as irrelevant answers. However, I believe that as time goes by, generative AI will become more and more useful, and its role in the e-commerce industry will become more and more important. McKinsey, an international authoritative organization, has also stated that in the next few years, AI will bring an additional 20% of revenue to the e-commerce industry.

After all, AI is a tool that makes our work more convenient and improves efficiency. So, in my opinion, although it is important for a tool to be "easy to use", the "people" who use the tool are more important.

Amazon

AI

listing