The News
New research published by a team at the University of Massachusetts Amherst aims to shed light on how changes to social media platforms’ algorithms alter the spread of political content.
The research, which focuses on Meta’s Facebook, comes at a moment of intense scrutiny around the online spread of political misinformation ahead of November’s presidential election in the US. It found that in 2020, Facebook’s algorithm was successfully altered to mute political and potentially harmful posts, preventing their spread, but that later algorithm changes may have reversed those effects.
Meanwhile, a separate new poll conducted by Axios and The Harris Poll (not associated with Vice President Kamala Harris) found that more Americans are worried about politicians spreading misinformation than they are about the role of social media companies, deceptive AI, or even foreign governments.
“It used to be, we were worried about China or Russia, fake ads or Facebook. Now, no, it’s coming from the campaigns,” the CEO of The Harris Poll said. The poll also found about 70% of the people surveyed believed misinformation would play a role in the upcoming election, while about 80% believed it could influence the outcome of an election.
Meanwhile, local government officials speaking to CNBC voiced concerns about misinformation spreading on Facebook, which they said was worsened by layoffs in the trust and safety and customer service teams, as well as the platform deprioritizing news.
SIGNALS
Meta’s algorithm changes may alter how information spreads
Last year, a Meta-backed study from several US universities published in Science concluded that Facebook’s algorithm muted “political and untrustworthy content” in the months around the 2020 US election, and didn’t significantly affect user polarization. In a new paper published Thursday, a group of independent researchers found that while Meta may be able to effectively filter “untrustworthy” political content, frequent algorithm changes make it hard to come to definitive conclusions about what works. Meta has down-ranked political content since 2020, but as seen with a recent false Facebook post about Haitian immigrants in Springfield, Ohio, the network’s vast user base means politicized misinformation can still spread, CNBC noted, and could be amplified by high-profile politicians.
The big (so far) unmaterialized fear: AI
Meta hasn’t announced any planned changes to its algorithm for this election cycle, although it did require labeling of political ads generated with AI. With so many global elections coinciding this year, some experts have warned that AI could drive a “tech-enabled Armageddon,” as fighting AI-generated and distributed misinformation has become “a fractured and difficult undertaking.” But those fears may be overblown, at least for now, the MIT Technology Review noted: “Our obsession with AI’s supposed impact is unlikely to make things better — and could even make them worse when it leads us to focus solely on the shiny new thing while distracting us from the more lasting problems that imperil democracies around the world.”
World governments take proactive approach to big tech
Some world governments have taken a markedly more interventionist tack to curb social media companies than the US’ piecemeal approach. Most recently, Brazil’s judiciary banned Elon Musk’s X in the country, which some experts consider to have been successful in fighting misinformation. The European Union, meanwhile, has passed sweeping antitrust laws targeting big tech, as well as legislation to ensure more transparency around ads and algorithms. However, critics warn so-called “fake news” laws could have chilling effects on press freedom in some cases, the Center for International Media Assistance noted.