Follow Slashdot blog updates by subscribing to our blog RSS feed

 


Forgot your password?
Close
typodupeerror
×
News

Brazilian Academics Create Automated Fake News Detection Platform (zdnet.com) 35

Posted by BeauHD from the real-vs-fake dept.
An anonymous reader quotes a report from ZDNet: A group of Brazilian researchers has created a web platform that is able to identify false information online in an automated manner. Developed by academics at the Center for Mathematical Sciences Applied to Industry (CeMEAI), the system uses a combination of statistical models and machine learning techniques to establish whether a specific content in Brazilian Portuguese is likely to be false. Initial tests suggest the platform is able to detect fake news with a 96% accuracy. The CeMEAI is a research center based in the mathematics and computer science department of the University of Sao Paulo, in the Sao Paulo state city of Sao Carlos. The center is supported by grants from the Sao Paulo Research Agency (FAPESP). In an interview with FAPESP's news agency, project coordinator and technology transfer director Francisco Louzada Neto said the goal of the project is "to offer society an additional tool to identify, not only subjectively, whether a news item is false or not."

The system uses statistical methods to analyze writing characteristics, such as words used or more frequently used grammatical classes. These are then fed into a machine learning-based classifier, which is able to distinguish patterns of language, vocabulary and semantics of fake and real news, and automatically infer whether the content submitted to the platform is false. The models were trained with a massive database of real and false news and were exposed to the vocabulary used in over 100,000 articles published over the last five years. The researchers will aim to use the false news related to the upcoming presidential elections, as well as content related to the Covid-19 pandemic to further calibrate the models. The researchers also commented on the potential risks of the system in the interview, including the potential that the system could be used by fake news creators to assess the potential for false content to pass for real before it is published. "That's a risk we're going to have to deal with," Louzada noted.
This discussion has been archived. No new comments can be posted.

Brazilian Academics Create Automated Fake News Detection Platform More

Brazilian Academics Create Automated Fake News Detection Platform

Comments Filter:

  • A good test... (Score:2, Insightful)

    by Anonymous Coward
    would be to feed it all the stories from the mainstream media about the "Trump Russian Collusion" story and see if it identifies it as 100% fake news.

    If it doesn't it's just another propaganda spreader.

    • Yeah, I'm sure those were all first written in Brazilian Portuguese.
    • Im not saying there was active collusion but it's quite obvious Trump rides the Putin peen.

    • Trump's campaign chairman was an unregistered foreign agent and went to jail for that, and Trump pardoned him.

      Paul Manafort lobbied for the former president of Ukraine, Viktor Yanukovych, and lied and hid it, while running a campaign for US president. That was the Ukrainian President who's fall brings us to Russia declaring war on Ukraine today.

      Great guy. And as the FBI director at the time said, Trump wasn't being investigated, but you bet your ass anything Manafort touched was.

      So, pardoned Manafort, and

    • Re: (Score:2)

      by kunwon1 ( 795332 )
      Not insightful, this is gaslighting.

      Collusion between Trump and Russia has been proven beyond a reasonable doubt

      Read the Mueller report

  • The real test for a fake news detector is an article that is specifically written to debunk fake news. As such an article will repeat many of the same points as the fake article, but explain why it's not correct, it would be easy to detect it as another fake article.

  • Stop believing headlines (Score:3)

    by Okian Warrior ( 537106 ) on Saturday February 26, 2022 @11:05AM (#62306147) Homepage Journal

    There's a lot of information on fake news in the scientific literature... that's total crap.

    Reading one paper, an actual scientific paper peer reviewed and published, was tracking the path of fake news echoing and their patterns: single points do most of the echoing, and if you could shut down three or four single points you could turn the volume down on fakeness.

    The paper used data from another paper, again peer reviewed and published, that is a database of vetted fake news articles.

    That paper outlined the collection methods, criteria, and other experimental design details and had a "representative sample" of links that it considered fake.

    I chose one link that looked interesting (to verify), it was an infowars link so Alex Jones, and let's go see what it says and...

    It was the original "spirit cooking" article from the time of the Hillary Clinton campaign, and the article was correct in all particulars. I could not find even a single statement in that article that was exaggerated or false. Nothing. Nada. The controversy *surrounding* the article had been amplified to the stars, but the original article was completely accurate.

    We're used to seeing headlines posted here on Slashdot, or reported in the news, and commenting as if the headlines were true, but when you drill down to base-line data you find that the actual situation is much different and far more sedate than the headline would suggest. Climate change articles are a good example, where a sane and calm scientific observation about something gets blown up out of proportion in an attempt to drive policy.

    It's the "telephone game" in reporting, where each news report puts just a little bit more spin on the original base line data.

    So in response to the OP, I'm completely doubtful that a Slashdot headline about a news report of an interview with the head of a news organization is going to be at all useful or accurate, and I might wonder about the (from the article) "massive database of real and false news" is accurate. In the past year alone, the Chinese lab leak theory went from "fake news, will get you banned" to "possibly true, here's the evidence". Is the "massive database" ever updated or corrected?

    Stop believing headlines, especially the ones you see on Slashdot.

    If you want to form an opinion, at the very least drill down to the scientific paper and read the abstract.

    • Re: (Score:2, Interesting)

      by kick6 ( 1081615 )

      There's a lot of information on fake news in the scientific literature... that's total crap.

      Reading one paper, an actual scientific paper peer reviewed and published, was tracking the path of fake news echoing and their patterns: single points do most of the echoing, and if you could shut down three or four single points you could turn the volume down on fakeness.

      The paper used data from another paper, again peer reviewed and published, that is a database of vetted fake news articles.

      You managed to invoke the god of peer review twice in the first 3ish paragraphs of your post. The only problem is: peer review is absolutely fucking broken. Your appeal to it's authority falls absolutely flat. There are no peers, nobody actually does reviews, it's more like grammar/spell check if you can actually get anyone to review your work, and a lot of actual good science is gatekept from being published simply because it disagrees with """scientific""" orthodoxy.

      Find another god to backstop your

  • Just follow a style guide like the The New York Times Manual of Style and Usage. [wikipedia.org]

  • Amateurs trying to play a pro game.

  • The researchers also commented on the potential risks of the system in the interview, including the potential that the system could be used by fake news creators to assess the potential for false content to pass for real before it is published. "That's a risk we're going to have to deal with," Louzada noted.

    So the concerns that it would be used to take down legitimate news (by accident or intentionally) is not a concern, huh?

  • This is similar to the problem of identifying spam. Once an algorithm is devised to detect it, spammers start tuning their approach to circumvent it. Ever wonder why you receive spam that seems utterly pointless or nonsensical? It could be that spammers are probing to see how filters react. It's like a little arms race, where improvements in detection are followed by improvements in circumvention, ad infinitum.

    • I believe the best usage for such tool is to assist detection of fake news, not automate it entirely. It would detect the most "amateurish" fake news and leave the "pros" ones to human eyes.

  • The description says it doesn't do fact checking or any content verification, but focuses on writing style and language use. This means that it's probably just learning how to recognize certain authors, or even authors from a certain ethnic, cultural, or otherwise associated group of people.

    I would say find some authors of the fake news on which the system was trained, have them write a non-fake news article and see the results. If the AI says >50% likely to be fake, that tells you it simply learned

  • the system uses a combination of statistical models and machine learning techniques to establish whether a specific content in Brazilian Portuguese is likely to be false.

    So it decides what's fake based on what it has seen before. The result is an Echo Chamber that nothing new can enter.

    Enjoy a worldview where everything new would be classified as fake news.

  • Given who runs Brazil currently, [bbc.com] I'd say any publicly-supported institution researching the sorting of fact from fiction in political discourse is going to be somewhat less than trustworthy in its results.

Slashdot Top Deals

Some people manage by the book, even though they don't know who wrote the book or even what book.

Close