Researchers at Sungkyunkwan University in South Korea said commonly used deepfake-generating methods can fool facial recognition services, VentureBeat reports. Using artificial intelligence models trained on five datasets containing faces of celebrities and politicians, the researchers created 8,119 manipulated videos and said application programming interfaces from Microsoft and Amazon were susceptible to attacks. “We believe our research findings can shed light on better designing robust web-based APIs, as well as appropriate defense mechanisms, which are urgently needed to fight against malicious use of deepfakes,” the researchers said.
8 March 2021
Study warns deepfakes can fool facial recognition services
Related stories
The increasing need to address digital governance
Notes from the IAPP Canada: Consultation period on CBPR implementation ends 30 June
CPPA Board tees up new consultation on draft ADMT, cybersecurity audit, risk assessment regulations
Ireland's DPC finalizes TikTok decision with potential data transfer ban
A view from DC: Colorado considers re-forging AI guardrails