Mark Zuckerberg may insist that Facebook doesn't have a serious fake news problem, but his company is apparently taking steps to identify misleading stories.

Some Facebook users tweeted Monday that they received surveys from the site asking them to identify whether certain headlines are misleading.

Chris Krewson, the editor of Philadelphia news outlet Billy Penn, noticed the query under a Philadelphia Inquirer article. It asked him to identify to what extent the link's title uses "misleading language," with options ranging from "not at all" to "completely."

Facebook is asking whether this @PhillyInquirer headline is fake? pic.twitter.com/cCUpwtvQlS

— Chris Krewson (@ckrewson) December 5, 2016

And writer Jorge Camargo got a survey with slightly different wording under a Rolling Stone article. That particular example, which asks whether the article's headline "withholds key details of the story," might be aimed at battling clickbait rather than so-called fake news, however.

A Facebook survey to see how accurate a Rolling Stone headline is. Pizzagate shows that information on social media fucking matters. pic.twitter.com/i4PIsbFhYF

— Jorge (@iamjorgecamargo) December 5, 2016

We haven't received either survey ourselves, and it's probably limited to a small subset of users. But Facebook confirmed to Mashablethat it is an official feature.

We asked whether this was a measure to combat fake news and how users are selected to receive the surveys, but the company did not share any details.

Fake news is a problem with real-world consequences. Over the weekend, a man armed with an assault rifle stormed into a restaurant in Washington, D.C. to "investigate" the #Pizzagate conspiracy theory. That theory, which alleges a bogus connection between Hillary Clinton and an invented child sex ring run out of Comet Ping Pong pizzeria, is clearly fake but has been shared extensively on social media.

For his part, Zuckerberg has said Facebook is looking to implement "better technical systems" to detect fake news. He's also mentioned several other approaches, including labelling stories as fake, using third-party verification services and listening to input from media professionals.

Asking for user input is a double-edged sword, though. On one hand, if users were able to easily detect fake news, Facebook wouldn't be in this mess. On the other, Facebook can probably compare this data to the other approaches it's using and see how well it performs before implementing human news curation as a feature.