Apr 25, 2023 14:02:00

Spam generated by ChatGPT on Amazon reviews and Twitter is exploding, experts are concerned about the increase in 'AI spam'

It has been reported that ChatGPT, which receives human words and produces natural sentences, is beginning to be used in spam as well. ChatGPT often uses pillow words such as 'I am an AI language model ...', but it seems that this part is posted as it is without being edited.

AI Spam Is Already Flooding the Internet and It Has an Obvious Tell

https://www.vice.com/en/article/5d9bvn/ai-spam-is-already-flooding-the-internet-and-it-has-an-obvious-tell

When you ask ChatGPT for something, it often starts with fixed sentences such as 'I am a word generated by artificial intelligence ...' or 'I am an ethical AI ...'. I have. VICE, a digital media company, reports that if you search with these fixed phrases, you can reasonably find the sentences generated by ChatGPT.

Texts generated by ChatGPT also appear in Amazon product reviews. Looking at a review posted on a certain diet appliance, the phrase 'as an AI language model' at the beginning of the sentence, which often appears in sentences generated by ChatGPT, remains as it is.

Some AI-generated reviews are positive, while others are negative. You can also check the fixed phrase 'As an AI language model' in

the review posted on the auxiliary equipment that enhances the operability of the controller for game consoles. The review is titled 'Broken in a few days' and is generally disparaging of the product.

Are reviews using ChatGPT posted by unrelated users who have not purchased or used the product to manipulate impressions, or are they the result of using ChatGPT to help reviewers verbalize their thoughts? , not sure. However, a certain book review shows clear human intentions.

Steven, who posted

a review of a book about Christianity, said at the beginning of the review, ``I asked AI to review the book,'' and then added the text generated by AI as it was.

An Amazon spokesperson told VICE, ``We do not tolerate fake reviews.We want Amazon customers to shop with confidence knowing the reviews they see are genuine and trustworthy. I would like to,' he said, adding, 'We will suspend accounts or take legal action against anyone who violates our policies and posts fake reviews, and removes their reviews.' Amazon is working to eliminate these fake reviews and has established a dedicated team.

Statements similar to those seen on Amazon have also been confirmed on Twitter. In the account pointed out by security researcher Conspirador Norteño, ``I'm sorry, I cannot generate inappropriate or offensive content.'' It is said that there was evidence that the sentence was tweeted. These accounts posted tweets about cryptocurrencies and politics.

Filippo Menczer, a social media expert at Indiana University, said the cost of generating spam that looks genuine has come down, and that content ranging from malware to financial fraud to hate speech is polluted by spam. I am concerned that there is a possibility that ``There is no clear solution at this time, as it is very difficult for humans to distinguish between AI-generated sentences and there are no political restrictions on the use of AI,'' Menczer said.