AI-made articles published on the news site ``CNET'' still contained serious errors even after human fact-checking

It was reported on the IT news site

CNET that AI was used to create financial commentary articles from around November 2022, but despite fact-checking by a human editor, the article The news site Futurism reports that an error was found in .

CNET's Article-Writing AI Is Already Publishing Very Dumb Errors

CNET used AI to write articles. It was a journalistic disaster. - The Washington Post

CNET Is Experimenting With an AI Assist.

On January 11, 2023, Futurism and web marketer Gael Breton discovered that a CNET financial article was written by an AI.

It is reported that the news media CNET has been secretly creating articles with AI with the signature ``CNET Money'' since November 2022-GIGAZINE

Later, on January 12, 2023, CNET admitted to using AI to create its financial articles, stating that all AI-generated articles were reviewed, fact-checked, and edited by a human editorial staff. announced .

Despite this, multiple mistakes have been found in articles created by AI. In the article 'What is Compound Interest?' published by CNET on January 12, 2023, 'If you deposit $3 in a savings account at a compound interest rate of 10,000% annually, you will earn $10,300 in interest at the end of the first year. ,' but the actual interest you will earn is $300.

The article was later revised , but Futurism pointed out, ``It highlights the technical challenges of sentence generation AI that creates sentences that seem to be accurate, regardless of the factual accuracy.'' I'm here.

Hany Farid of the University of California, Berkeley said, ``The AI-generated texts appeared authoritative, and editors let their guard down, resulting in less careful fact-checking than human journalists. I don't think so,' he said.

In a statement released on January 12, CNET editor Connie Guglielmo said CNET's use of AI was not to replace reporters with AI, but to assist humans in their work. It is called an 'experiment' for the purpose of ``Our goal is to see if AI can help busy reporters and editors who need to cover every topic,'' Guglielmo said.

CNET said on January 17, 2023, ``Because both humans and AI make mistakes, we actively check all AI-created works through the editing process to make sure there are no more mistakes.'' increase. In addition, 'We will continue to make necessary corrections even if errors are discovered again.'

At the time of writing the article, the statement ``This article was assisted by the AI engine and was reviewed, fact-checked, and edited by the editorial staff'' was added to the top of almost all articles published by CNET's AI. It is easy to judge that it is an article created by.


Stop publishing articles because CNET is writing articles with AI as a problem - GIGAZINE

in Web Service, Posted by log1r_ut