Microsoft announces a new tool to identify deepfakes and fight disinformation
By Abhishek Baxi · Sep 1, 2020
By Abhishek Baxi · Sep 1, 2020
[SHOWTOGROUPS=4,20,22]
Microsoft Для просмотра ссылки Войдиили Зарегистрируйсяa new technology to fight deepfakes that can be used to spread disinformation, especially in the upcoming US elections. Deepfakes, or synthetic media, use artificial intelligence to alter photos, videos, or audio files in hard-to-detect ways to make someone appear to do or say something they didn’t or to be in places they weren’t. These campaigns are intended to defame notable people or polarize the electorate.
Microsoft Video Authenticator analyses a still photo or video to provide a percentage chance, or confidence score, indicating the chance that they are manipulated. For videos, the tool can provide this score in real-time on each frame as the video plays. This technology was originally developed by Microsoft Research in coordination with Microsoft’s Responsible AI team and the Microsoft AI, Ethics and Effects in Engineering and Research (AETHER) Committee.
Microsoft Video Authenticator works by ‘detecting the blending boundary of the deepfake and subtle fading or greyscale elements’ that are often not detectable by the human eye.
The technology isn’t released to the public directly though. Microsoft is partnering with the AI Foundation, a dual commercial and nonprofit enterprise based in San Francisco. Video Authenticator will initially be available only through the foundation’s Reality Defender 2020 (RD2020) initiative to organizations involved in the democratic process, including news outlets and political campaigns.
One limitation of the technology is that it risks becoming quickly outdated as deepfake technology will continue to grow in sophistication. Microsoft has announced a separate system that will enable content producers to add hidden code to their footage so any manipulation can be flagged. Microsoft has also partnered with a consortium of media companies like the BBC, CBC/Radio-Canada, and the New York Times to test the technology and help advance it as a standard.
This initiative to combat disinformation is part of Microsoft’s Defending Democracy Program which also includes technologies like ElectionGuard to protect voting as well as AccountGuard to help secure campaigns.
[/SHOWTOGROUPS]
Microsoft Для просмотра ссылки Войди
Microsoft Video Authenticator analyses a still photo or video to provide a percentage chance, or confidence score, indicating the chance that they are manipulated. For videos, the tool can provide this score in real-time on each frame as the video plays. This technology was originally developed by Microsoft Research in coordination with Microsoft’s Responsible AI team and the Microsoft AI, Ethics and Effects in Engineering and Research (AETHER) Committee.
Microsoft Video Authenticator works by ‘detecting the blending boundary of the deepfake and subtle fading or greyscale elements’ that are often not detectable by the human eye.
The technology isn’t released to the public directly though. Microsoft is partnering with the AI Foundation, a dual commercial and nonprofit enterprise based in San Francisco. Video Authenticator will initially be available only through the foundation’s Reality Defender 2020 (RD2020) initiative to organizations involved in the democratic process, including news outlets and political campaigns.
One limitation of the technology is that it risks becoming quickly outdated as deepfake technology will continue to grow in sophistication. Microsoft has announced a separate system that will enable content producers to add hidden code to their footage so any manipulation can be flagged. Microsoft has also partnered with a consortium of media companies like the BBC, CBC/Radio-Canada, and the New York Times to test the technology and help advance it as a standard.
This initiative to combat disinformation is part of Microsoft’s Defending Democracy Program which also includes technologies like ElectionGuard to protect voting as well as AccountGuard to help secure campaigns.
[/SHOWTOGROUPS]
