Published 19:39 IST, September 6th 2019
Facebook, Microsoft partner to detect deepfake videos through contest
Deepfakes challenge: Facebook, Microsoft and seven universities, including Oxford and MIT, have come together to launch a deepfake challenge. Details inside
Advertisement
Facebook, Microsoft and seven universities, including Oxford and MIT, have come toger to launch a "deepfake challenge." Toger, y are looking to fight Deepfakes and improve tools that detect videos and or media manipulated by artificial intelligence (AI). Facebook on Thursday anunced its 10-million-dollar initiative. Facebook aims to curb what is seen as a significant threat to integrity of online information, including on social media services like Facebook and Instagram. As part of initiative, Facebook has partnered with Microsoft for AI expertise. Facebook has also partnered with acemics from Massachusetts Institute of Techlogy (MIT), Cornell University, University of Oxford, University of California-Berkeley, University of Maryland and University at Albany. se partnerships between Techlogy firms and universities represent an effort to combat dissemination of manipulated video or audio as part of a misinformation campaign.
Advertisement
industry-backed Partnership on AI
" goal of challenge is to produce techlogy that everyone can use to better detect when AI has been used to alter a video in order to misle viewer," said Facebook chief technical officer Mike Schroepfer. Schroepfer said deepfakes techniques "have significant implications for determining legitimacy of information presented online. Yet industry doesn't have a great data set or benchmark for detecting m."
What is Deepfake?
What is deepfake, you may ask? Deepfakes present realistic generated videos of people doing and saying fictional things, using Artificial Intelligence (AI). Earlier, a deepfake video of Facebook CEO Mark Zuckerberg was uploed on Instagram. In video, Facebook CEO was portrayed as saying things that seemed to defame Facebook and Zuckerberg himself. However, Instagram refused to delete video in question. Currently, Facebook is evaluating how to handle “deepfake” videos created with artificial intelligence (AI) and high-tech software tools to produce a false but realistic clip
Advertisement
Deepfake challenge
This happens to first project of a committee on AI and media integrity. Created by Partnership on AI, mission of project is to promote beneficial uses of artificial intelligence and is also backed by Apple, Amazon, IBM and or tech firms and n-governmental organisations. Terah Lyons, executive director of Partnership, said new project is part of an effort to stem AI-generated fakes, which "have significant, global implications for legitimacy of information online, quality of public discourse, safeguarding of human rights and civil liberties, and health of democratic institutions". Facebook said it was offering funds for research collaborations and prizes for challenge, and would also enter competition, but t accept any of prize money.
Advertisement
Oxford professor Philip Torr, one of acemics participating, said new tools are "urgently needed to detect se s of manipulated media. "Manipulated media being put out on internet, to create bogus conspiracy ories and to manipulate people for political gain, is becoming an issue of global importance, as it is a fundamental threat to democracy," Torr said in a statement.
Advertisement
17:53 IST, September 6th 2019