Deepfake expert Hany Farid warns AI‑generated media threatens elections and finance

Artificial‑intelligence‑generated deepfake videos have surged from about 500,000 in 2023 to roughly 8 million last year, according to the DeepStrike monitor. In the United States, campaigns and outside groups are already using synthetic videos to portray candidates in wildly inaccurate scenarios ahead of the 2026 midterms.

Hany Farid, a leading digital‑forensics scholar, told the New York Times that the quality of the fakes has made his own visual analysis nearly impossible, saying "within a year or two, our whole visual system will be utterly useless." He expressed personal doubt about his ability to detect fakes and announced plans to leave Silicon Valley for a farm in rural Vermont.

Beyond political manipulation, the reports note that deepfakes are being weaponised for financial fraud. Synthetic borrowers—AI‑generated personas that can navigate loan‑underwriting checks—are emerging, challenging banks and fintech firms that rely on data to assure certainty. Detecting such fraud is costly and time‑consuming, while the speed of social‑media diffusion means a fake can become accepted fact within minutes.