University researchers say deepfakes are increasingly plausible and threaten shared facts

Dr. Alex Scott, assistant professor in the University of Iowa School of Journalism and Mass Communication, told attendees that synthetic visual media and deepfakes are proliferating and that many popular generators now produce convincing audio-visual content. He played a Jordan Peele–produced PSA to demonstrate how voice and image synthesis can convincingly simulate public figures and noted study results showing rising susceptibility: experiments cited in the talk found deception rates climbed markedly between 2022 and 2025.

Scott described political uses of synthetic media—manipulated clips used to ridicule or misrepresent opponents, slowed videos that suggest impairment, and AI-generated ads that evade straightforward labels—and warned that rapid, incidental exposure on social platforms can make verification difficult. He outlined several responses: media literacy and inoculation training, platform labeling and moderation, government measures such as statutes addressing non‑consensual intimate-image deepfakes and proposed political-use restrictions, detection tools, and content-provenance systems like Adobe’s Content Authenticity Initiative, while noting each approach has limits in practice: “If you’re trying to deceive somebody, you’re not putting in a content credential,” he said.

Scott urged a combined strategy—education, policy, technical tools and public pressure—to preserve verifiable and accountable public communication.