(Credit: Koshiro K/Shutterstock)
Like TikTok, Meta has announced plans to allocate resources to finding and identifying misinformation across its platforms ahead of the European Union's Parliament elections in June.
Meta is establishing an Elections Operations Center to tackle the problem of fake news, says Head of EU Affairs Marco Pancini. He adds that Meta created an election integrity team last year, but the EU operations group is set to launch "as the election approaches."
Meta says it will remove any content from its platforms—which includes Facebook, Instagram, and Threads—that could discourage voters from taking to the polls or possibly incite "imminent violence or physical harm."
Any other content that could possibly fall under misinformation will be passed on to one or more of Meta's 26 fact-checking partners, which will assess the content and label it if it's fake or misleading, but won't remove it. The company claims that 95% of users shown labeled posts don't click through to view the fake information anyway. Instead, Meta says it will suppress posts found to be spreading fake info so that fewer users will see them in their feeds overall.
The tech giant seems to have stricter rules for ads, though. Meta won't allow any fake news to appear in ads at all, according to Pancini's post. It will also take down ads that discourage voting, spread news of fake wins or losses, or encourage conspiracies surrounding the EU election's legitimacy, for example.
When it comes to AI, Meta announced its AI labeling policy earlier this month that offers early stage labeling for content made with its own AI tool and other AI content that's already been given a hidden "metadata" label. This policy doesn't address malicious deepfakes where creators intentionally remove this metadata, though. Both Meta and OpenAI have admitted that it's not hard to remove such data, either.
But Meta plans to add the ability for users to self-identify their content as AI-generated, and could punish those who fail to add the label to their AI creations. The company hasn't yet shared when such a labeling tool will be launched, or what the penalty might be for posting unlabeled AI-generated content.
"If we determine that digitally created or altered image, video or audio content creates a particularly high risk of materially deceiving the public on a matter of importance, we may add a more prominent label, so people have more information and context," Pancini says of Meta's two-tiered AI content labeling plans.
But not all misleading media is AI-generated. Sometimes, like in the case of the President Biden video Meta refused to take down, content is simply edited in a way that deceives viewers. Meta claims it will label content as "altered" if it's been "manipulated or transformed" in any way that doesn't necessarily involve the use of AI. Like with fake news posts, Meta plans to similarly suppress any content with the "altered" label across its news feed algorithms.
Meta previously blocked any Russian state-controlled media outlets from posting to EU audiences, and labels any media accounts owned or funded by national governments. Citing tech firm Graphika's data, Meta says that since it implemented the policy, "more than half of all Russian state media assets had stopped posting altogether" on Facebook and Instagram as of January this year.
Russia itself has taken a strong stance against Meta, calling for the arrest of the company's communications director for alleged "extremist activities." But Meta employees have said the company is simply "focused on protecting people's right to speech as an expression of self-defense" regarding Russia's invasion of Ukraine.