June 15, 2020 4:03:19 pm
Facebook announced the winners of its $1 million “Deepfake Detection Challenge”, which is exactly what it sounds like— a contest for AI models to detect doctored deepfake videos that “convincingly” imitate real people. However, as the winning model scores 65.18 per cent accuracy rate, it looks like there is much work that needs to be done before we rely on such algorithms to spot deepfake content.
With over 2,000 participants, the competition saw more than 35,000 AI models attempting to detect deepfakes from a black box dataset of 100,000 videos. Facebook reportedly hired more than 3,000 actors to create these video clips where they were recorded holding conversations in natural environments.
Also read | Explained: What is deepfake?
Some of these clips were altered by having other actors’ faces pasted on their’s using AI. Participants were given access to this data to train their algorithms. When the AI models were tested on this material, they produced accuracy rates as high as 82.56 per cent. However, when the same algorithms were tested against a “black box” dataset of unseen footage, their performance dropped significantly with the best-scoring model achieving an accuracy rate of 65.18 per cent.
Concerns regarding deepfake videos
There are growing concerns regarding the misuse of deepfakes, which can put words in anyone’s mouth as evident in the deepfake video made by Buzzfeed. Here, the words of comedian, writer, and director Jordan Peele’s words are imposed on the then US President Obama:
Social media apps had banned deepfake videos on their platforms. At the moment, while the political effects of deepfake videos have been minimal, its use in the creation of nonconsensual pornography has already started to make noise. Luckily, pornographic content is easy to identify and remove from the social media platforms.
Express Tech is now on Telegram. Click here to join our channel (@expresstechnology) and stay updated with the latest tech news
As per the reports, Facebook is currently developing its own deepfake detection technology that is separate from this competition. Fake news is already a menace for social media platforms, that is why they want to be ready with tools that can fish out deepfake videos when they hit them.
📣 The Indian Express is now on Telegram. Click here to join our channel (@indianexpress) and stay updated with the latest headlines
- The Indian Express website has been rated GREEN for its credibility and trustworthiness by Newsguard, a global service that rates news sources for their journalistic standards.