Study Reveals Miscontextualized Real Images Outpace Deepfakes in Multimodal Misinformation

A study analyzing 27,000 posts in seven languages on X introduces MMMMM, a taxonomy and dataset for multimodal misinformation. It finds that miscontextualized real images are more common than deepfakes, and image-text misinformation is more prevalent and harmful than text alone.

2026-09-01 ~ 2026-09-02 · 2 related posts