AI deepfakes of late‑night hosts Jimmy Kimmel and Jon Stewart spread online

Social media platforms are increasingly flooded with AI‑generated celebrity deepfakes, but experts warn that videos of late‑night television hosts such as Jimmy Kimmel and Jon Stewart are among the easiest to produce and could have a disproportionate impact. The combination of high‑quality, publicly available footage and the hosts’ distinctive voices and mannerisms allows automated tools to create convincing impersonations with relatively little data, a process that is faster and less resource‑intensive than deepfakes of other public figures.

The ease of creation raises concerns about misinformation and reputational harm. In a recent study, researchers found that deepfakes of Kimmel and Stewart were recognized as authentic by a majority of viewers even when the videos were clearly fabricated. The potential for these videos to be shared widely, coupled with the hosts’ influence on public discourse, means that such deepfakes could be used to spread false statements or satirical content that is mistaken for genuine commentary.

Industry watchdogs and platform moderators are calling for tighter verification protocols and clearer labeling of AI‑generated content. While the technology behind these deepfakes is advancing rapidly, the current trend underscores the need for both media literacy and robust safeguards to prevent the spread of deceptive videos that could mislead audiences and undermine trust in broadcast journalism.