Tech Platforms Put AI Between Voters and Facts

Artificial intelligence now sits inside every major social media platform ahead of the 2026 midterm elections, and company tools increasingly shape voters’ choices. Chatbots, media-generation services and automated moderation stand between people and election information, while the companies have responded to the risks inconsistently—or not at all.

Platforms Set the Terms

“We're in uncharted territory at this point,” said Deen Freelon, a professor of communication at the University of Pennsylvania who studies digital politics. His warning comes as technology companies have rapidly brought AI tools to platforms voters use, with some services arriving just months before the election.

The New York Times reported that this expansion may create obstacles for voters seeking reliable election information. Researchers said AI tools have contributed to biased and inaccurate information about polls and have amplified foreign influence operations. People who encounter election content bear the consequences, while platforms control the services that circulate it.

AI-generated material creates another problem. Researchers said low-quality “slop” has risen through the ranks, tainting the training data that informs chatbot responses. According to the research described in the report, poor-quality material can feed a cycle that affects what chatbots tell users. The report gives no figures for how many voters have encountered such content or how much it has changed anyone’s choices.

Questions Without Answers

Lawmakers and watchdog groups have asked technology companies to explain their testing, safeguards and the extent of voter interactions with AI. They say the companies haven't shared enough information. Those seeking accountability want answers, but the companies with information about their systems haven't provided enough, according to the groups.

That imbalance matters in an election shaped by privately operated platforms. The report describes no public process for voters to examine how companies test AI services, how safeguards work or how often people interact with the tools. It does say lawmakers and watchdog groups are asking those questions, but the companies haven't supplied the detail they seek.

The election itself offers no answer to that gap. The report describes AI’s influence on voter choices and information, but provides no evidence that casting a ballot will make platform companies disclose their testing or safeguards. People take part in politics through systems whose workings remain unclear to those who use them.

More Powerful Tools, Same Secrecy

During the 2024 national election, platforms offered basic AI image-generation tools. By 2026, they had expanded into advanced services that combine audio, video and image capabilities. That marks a substantial shift in what platforms can generate, even as the report says companies’ responses to the risks remain inconsistent or absent.

The report names Tiffany Hsu and Steven Lee Myers as reporting from San Francisco, and Julian E. Barnes as reporting from Washington. It identifies no grassroots organizing, mutual-aid effort or community response to the platforms’ AI systems. It does document pressure from lawmakers and watchdog groups, alongside a continuing lack of information from the companies. The tools keep expanding. Public answers haven't kept pace.