It is pointed out that the article creation AI used by the news media CNET ``plagiarized articles of writers of competitors and affiliated sites''
In January 2023, it was discovered that the news media CNET was secretly publishing a financial commentary article that used AI to create, and it caused a great deal of attention. CNET has already stopped publishing AI-made articles, but a new technology media Futurism reported that 'CNET's AI was plagiarizing from articles written by competitors and group company writers.'
CNET's AI Journalist Appears to Have Committed Extensive Plagiarism
Futurism reported in January 2023 that CNET had been using AI for some financial commentary articles from around November 2022. CNET's AI-made articles were signed with 'CNET Money', making it difficult to tell at first glance that they were articles by AI. In addition, it is clear that Red Ventures, the parent company of CNET, published AI-made articles not only on CNET but also on affiliated sites such as Bankrate and CreditCard.com.
It is reported that the news media CNET has been secretly creating articles with AI with the signature ``CNET Money'' since November 2022-GIGAZINE
A subsequent investigation reported that CNET's AI-produced article contained serious errors, even though it had undergone human fact-checking.
AI-made articles posted on the news site ``CNET'' still contained serious errors even after human fact-checking-GIGAZINE
Following a string of controversial reports, CNET executives have decided to stop publishing all AI-generated articles. CNET editor Connie Guglielmo said in a statement on January 12 that the use of AI was not intended to replace reporters with AI, but to assist humans in their work. We call it an 'experiment'.
Stop publishing articles because CNET is writing articles with AI as a problem - GIGAZINE
Newly reported by Futurism, CNET's AI is copying and paraphrasing verbatim articles written by 'human writers' on competitors and affiliated sites, all without proper credit. . For example, CNET's AI-made article about 'overdraft protection' that avoids 'overdraft fee', which is a fee when payment exceeds the deposit balance, It contains sentences such as:
How to avoid overdraft and NSF fees
Overdraft fees and NSF fees don't have to be a common consequence.
Japanese translation:
How to avoid overdraft fees and NSF (insufficient balance) fees
Overdraft fees or NSF fees do not have to be typical results. There are some steps to avoid them.
On the other hand, another article published in competitor Forbes Advisor has very similar sentences.
How to Avoid Overdraft and NSF Fees
Overdraft and NSF fees need not be the norm.
Japanese translation:
How to avoid overdraft fees and NSF (insufficient balance) fees
Overdraft fees and NSF fees don't have to be standard. To get around them, you have some tools at your disposal.
Comparing the text of the AI-produced article with the text of Forbes Advisor, you can see that the content is almost the same, although there are changes in capitalization, paraphrasing of wording, and changes in syntax.
Also, comparing the headline of another article written by CNET's AI and the article of Forbes is as follows.
CNET:
Can You Buy a Gift Card With a Credit Card?
Forbes:
Can You Buy Gift Cards With a Credit Card?
Both are headings that mean ``Can I buy a gift card with a credit card?'' The only difference between the two is ``Gift Card'' (singular) or ``Gift Cards'' (plural) am.
In addition, Futurism points out that CNET's AI also used the writings of writers on affiliated site Bankrate. A comparison of the CNET article and the Bankrate article is as follows.
CNET:
Becoming an authorized user can help you avoid applying for a card on your own, which is a major benefit if you currently have bad credit or no credit history.
Japanese translation:
By becoming an authenticated user, you do not need to apply for a card yourself, which is a great advantage for people with low current credit or no credit history.
Bank rate:
Becoming an authorized user also lets you avoid having to apply for a card on your own, which is a major benefit if you currently have bad credit or no credit history at all.
Japanese translation:
Being an Verified User eliminates the need to apply for a card yourself, which is a great advantage for those with low current credit or no credit history at all.
Futurism says, “Basically, CNET’s AI seems to approach a topic by looking at similar articles that have already been published and reading sentences from them. , and adjust the structure, sometimes mashing up two sentences, breaking one sentence apart, and assembling parts into new patchwork sentences. Upon closer inspection, it looks more like an automated plagiarism machine pumping out articles to fire human journalists rather than a sophisticated text generator.'
The reason why CNET's AI is like a plagiarism machine is that the machine learning system itself inputs a huge amount of 'training data' and allows advanced algorithms to perform similar work. Futurism argued.
Professor Jeff Schatten, who is investigating fraud using AI at the University of Washington and Lee in the United States, reviewed the results of CNET's AI plagiarism survey submitted by Futurism, and found that a series of articles reached the level of plagiarism. He replied that he was. Schatten told Futurism, ``If a student published something equivalent to something CNET produced in my class and didn't cite the source, I would definitely consider it plagiarism.'' .
Futurism Editor-in-Chief Jon Christian said, 'I think there should be at least a way for writers, artists, and other creators to opt out of being included in the learning data (and if technically they can, however they can. from the trained model)”.
Yeah, I do think that at the *very* least there needs to be a way for writers/artists/other creators to opt out of inclusion in training data (and hopefully trained models as well, whatever that would look like technologically)
—Jon Christian (@Jon_Christian) January 23, 2023
Related Posts:
in Software, Web Service, Posted by log1h_ik