Facebook is making its own AI deepfakes to head off a disinformation disaster
The CTO of Facebook says videos forged using AI will be used maliciously on its platforms before long, reports Will Knight.
The logic: As they become easier to produce and more convincing, Facebook fears deepfakes could be the next big source of viral misinformation. It’s directed its own team of AI researchers to produce a set of deepfake videos of actors doing and saying routine things, set to be released at the end of the year.
Why? The idea is that they will help researchers build and refine detection tools. Although such methods exist, they often involved painstaking expert analysis. Tools for catching deepfakes automatically are only just emerging.
A challenge: Facebook will spend $10 million on detection tech, launching theDeepfake Detection Challenge with Microsoft and academics from MIT, Oxford University, and other institutions.
Unintended consequences: Despite all the fears over the use of deepfakes in the political arena, the technology is more likely to become a potent tool of cyber-stalking and bullying in the short term, says Henry Ajder, an analyst at Deeptrace. Read the full story here.
Deep dive: For more on the emerging threat of deepfakes and where the technology is heading, check out our profile of Hao Li, the world’s top deepfake artist.