MIT Proposes Using Crowdsourcing to Avoid Fake News

While Facebook is not convinced by crowdsourcing to fight against ‘fake news’, a study by MIT says it is a useful tool, especially to discriminate against the most extreme media.

Share Give it a Spin!
Follow by Email

While Facebook is not convinced by crowdsourcing to fight against fake news, a study by MIT says it is a useful tool, especially to discriminate against the most extreme media.

Fake news has become a not too desired protagonist in the global news scene, with decisive implications in the public perception of politics, economy or health. It is not necessary to dig a lot into the recent past to find clear examples of it: the unexpected electoral victory of Donald Trump, the rise of populisms in Europe or the growing weight of unconscious anti-vaccines are just some of them.

That is why the media and organizations around the world are turning to possible formulas that can end disinformation and, at the same time, preserve the freedom of the press and of expression. A difficult combination to achieve but which is plausible if we join forces as a society.

Or at least that’s what scientists at MIT think, who propose using crowdsourcing to avoid false news on the Internet. In a groundbreaking study focused on the US, academics have shown that collective collaborative judgments about the quality of news sources can effectively filter out false news and other types of misinformation online.

“What we discovered is that, while there are real disagreements between Democrats and Republicans with respect to the mainstream media, basically everyone (Democrats, Republicans and professional fact-checkers) agree that you can not trust the false and hyperparty sites,” explains David Rand, an MIT expert and co-author of the study.

In fact, by using a couple of public opinion polls to evaluate 60 news sources, the researchers discovered that Democrats relied on traditional media more than Republicans, with the exception of Fox News, which Republicans trusted. much more than the Democrats. But when it comes to lesser-known sites that offer false information, as well as “hyper-partisan” political websites, both Democrats and Republicans show a similar indifference to such sources.

The confidence levels for these alternative sites were generally low. For example, in one of the surveys, when respondents were asked to give a confidence rating of 1 to 5 for the media, the result was that hyperparty websites received a confidence rating of only 1.8. Republicans and Democrats, while fake news sites received a confidence rating of only 1.7 from Republicans and 1.9 from Democrats. In contrast, the mainstream media received a confidence rating of 2.9 from Democrats, but only 2.3 from Republicans; Fox News, however, received a confidence rating of 3.2 from Republicans, compared to 2.4 for Democrats.

The study adds a new twist to a topic of high interest. Not in vain, false news has proliferated on the Internet in recent years, and social networking sites like Facebook have received strong criticism for giving them visibility. Facebook rejected in January 2018 to allow readers to rate the quality of the news sources they see. But the MIT study suggests that such a crowdsourcing approach could work well, if implemented correctly. “If the goal is to eliminate really bad content, this really looks pretty promising,” says Rand. The social network ended up implementing a different functionality in August of last year, in which instead of assessing the news sources, the user who promulgates them is qualified.

And is that, as we say, because respondents generally distrusted the most marginal websites, there is significant agreement between the general audience and the verifiers of professional facts. That in turn implies that the crowdsourcing approach could work especially well by marginalizing false news stories, for example, by constructing audience judgments using an algorithm that ranks stories by quality. However, crowdsourcing would probably be less effective if a social media site tried to build consensus on the best sources of news and stories.