Experts Warn of Widespread AI Disinformation on Social Media Ahead of 2026 US Midterms

Experts are sounding alarms over a surge of A.I.-driven disinformation flooding social media platforms ahead of the November U.S. midterm elections. They caution that many misleading posts are likely “slipping through the cracks” without adequate scrutiny.

As millions of Americans gear up to vote, social media companies have reduced efforts to shield users from misinformation intended to sway elections.

In the wake of the Cambridge Analytica scandal, many platforms had strengthened their trust and safety measures. However, a backlash from certain political factions—fueled by concerns over perceived restrictions on free speech since Elon Musk’s takeover of X (formerly Twitter) and Donald Trump’s return to power in 2024—has led to significant cuts in relevant teams.

The advent of artificial intelligence has simplified the process of executing disinformation campaigns. Platforms are increasingly depending on AI rather than human moderators to manage content issues.

Frances Haugen, a former data engineer at Facebook known for her whistleblowing efforts, noted, “You can run an information operation now with quality that was impossible in 2016 for very little cost.” She explained how disinformation operatives can navigate content moderation by continuously modifying posts, allowing many to bypass scrutiny.

Political campaigns, from Pennsylvania to Texas, are frantically contesting lies that are proliferating online. Unlike in previous election cycles, they face a daunting challenge: determining whom to consult when fabrications circulate widely.

Both Meta, the parent company of Facebook and Instagram, and X declined to comment on the situation. YouTube, owned by Google, stated that all users are subject to uniform guidelines.

Texas Senate candidate James Talarico recently took a break from campaigning to recover from the flu. However, social media falsely claimed he was receiving treatment at a mental health facility in Ohio. This baseless narrative was fueled by a right-wing provocateur who disseminated a misleadingly edited phone conversation with a hospital employee.

Talarico’s campaign countered this by providing video footage showing him at home in Texas and clarified that he had never traveled to Ohio. Nevertheless, the falsehood gained traction and persisted for several days.

The Digital Democracy Institute of the Americas (DDIA) monitored the misinformation’s spread, noting how it evolved as it was replicated in conservative Spanish-language media. What began as a viral rumor morphed into false claims surrounding Talarico’s mental health.

This incident exemplifies a continuous struggle in combating electoral disinformation, as DDIA researchers highlighted how misinformation tends to shift, rendering original corrections inadequate.

Roberta Braga, DDIA’s founder, remarked that social media platforms enforce rules more stringently in English than in Spanish. She noted, “There are significantly fewer safeguards to tackle Spanish-language disinformation, especially with recent cuts to content moderation and a shift towards community-driven moderation methods.”

Talarico’s Republican opponent, Ken Paxton, also released an advertisement employing AI-generated visuals of Talarico allegedly reinterpreting the Bible, a tactic Talarico condemned as damaging to civic discourse.

Disinformation tactics once relegated to fringe circles are now amplified by official government sources, as seen when misleading posts claimed an undocumented immigrant was a campaign worker for Democrat Janelle Stelson. This narrative was further echoed by a statement from the Department of Homeland Security.

Despite evidence to the contrary showing that the immigrant had no affiliation with the campaign, misleading ads have continued to circulate.

Though social media platforms maintain policies against misinformation, their definitions may prove overly narrow, often limiting enforcement to misinformation directly impacting voting logistics rather than broader political falsehoods.

Both Meta and YouTube have altered their policies, sometimes enabling false claims to flourish unchecked under the guise of protecting political discourse. Users of these platforms navigate an environment where deepfakes and misinformation persist despite existing safeguards.

As platforms increasingly lean on AI for content moderation, the efficacy of these systems raises serious concerns. Haugen emphasized the need to focus on the broader distribution systems of misinformation rather than solely on the content itself.