Cybersecurity firms are researching techniques to foil potential deepfake attacks, which doctor online content to disseminate false information, as the technology becomes increasingly sophisticated, simpler, and more affordable.
Some companies are employing artificial intelligence to spot inconsistencies in fake media, based on knowledge of deepfake generation—although a sufficiently large dataset for a detection model is currently lacking.
Darren Shou of Symantec said the company is exploring ways of mapping the provenance of video and audio media, as an indication of potential authenticity or fraudulence.
Symantec also is looking into repurposing three-dimensionally-printed glasses—designed to deceive facial-recognition software—to prevent deepfakes of targeted corporate executives.
Meanwhile, startups like ProofMode and Truepic provide solutions that assign watermarks to photos, to ensure trustworthiness.
From Financial Times
View Full Article - May Require Paid Subscription
Abstracts Copyright © 2019 SmithBucklin, Washington, DC, USA
No entries found