False News: Are Smart Bots the Answer?

November 7, 2019

To us, this comes as no surprise—Axios reports, “Machine Learning Can’t Flag False News, New Studies Show.” Writer Joe Uchill concisely summarizes some recent studies out of MIT that should quell any hope that machine learning will save us from fake news, at least any time soon. Though we have seen that AI can be great at generating readable articles from a few bits of info, mimicking human writers, and even detecting AI-generated stories, that does not mean they can tell the true from the false. These studies were performed by MIT doctoral student Tal Schuster and his team of researchers. Uchill writes:

“Many automated fact-checking systems are trained using a database of true statements called Fact Extraction and Verification (FEVER). In one study, Schuster and team showed that machine learning-taught fact-checking systems struggled to handle negative statements (‘Greg never said his car wasn’t blue’) even when they would know the positive statement was true (‘Greg says his car is blue’). The problem, say the researchers, is that the database is filled with human bias. The people who created FEVER tended to write their false entries as negative statements and their true statements as positive statements — so the computers learned to rate sentences with negative statements as false. That means the systems were solving a much easier problem than detecting fake news. ‘If you create for yourself an easy target, you can win at that target,’ said MIT professor Regina Barzilay. ‘But it still doesn’t bring you any closer to separating fake news from real news.’”

Indeed. Another of Schuster’s studies demonstrates that algorithms can usually detect text written by their kin. We’re reminded, however, that just because an article is machine written does not in itself mean it is false. In fact, he notes, text bots are now being used to adapt legit stories to different audiences or to generate articles from statistics. It looks like we will just have to keep verifying articles with multiple trusted sources before we believe them. Imagine that.

Cynthia Murrell, November 7, 2019

Comments

Got something to say?





  • Archives

  • Recent Posts

  • Meta