Enjoy Free Shipping on Orders Above US$39.99! Learn More About Our Shipping Policy

Lazy use of AI leads to Amazon products saying ‘we can’t fulfill that request’


We know naming new products can be difficult, but these Amazon sellers chose some particularly strange names.
Expanding / We know naming new products can be difficult, but these Amazon sellers chose some particularly strange names.

Amazon

At this point, Amazon users are used to search results filled with counterfeit, fraudulent, or literally trash products. But these days, you may also have to choose products that are obviously questionable, such as, “Sorry, we can’t fulfill your request because it violates OpenAI’s usage policy.”

As of this writing, some version of that apparent OpenAI error message is appearing on a variety of Amazon products, from lawn chairs to office furniture to Chinese religious tracts (UPDATE: original published shortly after) (the link now goes to an archived copy). Several similarly named products that were available as of this morning were removed as information on the listing spread across social media (one such example is archived here).

Pro tip: Don't ask OpenAI to integrate trademarked brand names when generating names for oddly long rubber tubes.
Expanding / Pro tip: Don’t ask OpenAI to integrate trademarked brand names when generating names for oddly long rubber tubes.

Other Amazon product names don’t specifically mention OpenAI, but they clearly do, such as “Sorry, we can’t generate a response for your request” or “We’re sorry, we can’t provide the information you’re looking for.” I’m seeing an AI-related error message. ‘ (available in various colors). In some cases, the product name may even highlight the specific reason why an obvious AI generation request failed, and OpenAI may It points out that it is not possible to provide “advertised” content, or in some cases “. Encourage unethical behavior. ”

If you repeatedly call
Expanding / It’s especially ironic that this description repeatedly cites a “commitment to providing trusted product descriptions.”

The descriptions of these oddly named products are filled with obvious AI error messages like, “Sorry, we can’t provide the information you’re looking for.” The product description for the table and chair set (now removed) cheerfully read: It can be used for various tasks such as. [task 1], [task 2]and [task 3]Another product description (archive link) that appears to be for a tattoo ink gun states that “We prioritize accuracy and reliability by providing only verified product details to our customers.” He repeatedly apologized and said he could not provide any further information.

spam spam spam spam

It is not against Amazon’s policies to use large language models to generate product names and descriptions. On the contrary, in September, Amazon launched its own generative AI tool that allows sellers to “create more thorough and engaging product descriptions, titles, and product details.” And as of this writing, we’ve only found a handful of Amazon products that slip by with obvious error messages in their names or descriptions.

Still, these error message-filled listings highlight the lack of care and even basic editing that many Amazon scammers take when posting spam product listings on Amazon Marketplace. While some sellers can easily get caught posting OpenAI errors, there are probably countless others who use this technology to create product names and descriptions. Seem As if it were written by someone with actual experience using the product in question.

A set of apparently real people talking on Twitter/X.
Expanding / A set of apparently real people talking on Twitter/X.

Amazon isn’t the only online platform where these AI bots are deploying themselves. A quick search for “against OpenAI policy” or “as an AI language model” will turn up many artificial posts. Twitter / × For example, Threads or LinkedIn.security engineer dan feldman Amazon also pointed out a similar issue in April.However, searching for the phrase “as an AI language model” doesn’t seem to produce results that are clearly AI-generated these days.

While it’s fun for AI-generated content producers to point out obvious mishaps like these, a flood of hard-to-detect AI content is popping up in everything from art communities to science fiction magazines to Amazon’s e-book marketplace. may overwhelm you. Nearly every platform that accepts user submissions, including text and visual art, is now worried about a wave of AI-generated work crowding out the human communities it was created for. need to do it. It’s a problem that’s likely to get worse before it gets better.

[Update: In a statement provided to Ars Technica, Amazon spokesperson Maria Boschetti said that “We work hard to provide a trustworthy shopping experience for customers, including requiring third-party sellers to provide accurate, informative product listings. We have removed the listings in question and are further enhancing our systems.”]

Listing image by Getty Images | Leon Neal





Source link

We will be happy to hear your thoughts

Leave a reply

TCS Marketplace
Logo
Compare items
  • Total (0)
Compare
0
Shopping cart