SAN FRANCISCO--(BUSINESS WIRE)--The AI Foundation Nonprofit, in collaboration with top artificial intelligence (AI) researchers, industry partners, campaigns, news agencies and social media platforms, and pioneers of deepfake detection, is launching Reality Defender 2020 to help protect the 2020 elections from the threat of deepfakes. Reality Defender 2020 is part of AI Foundation’s larger Reality Defender initiative of synthetic media detection, specifically geared towards political disinformation and manipulation.
Reality Defender 2020 is a non-partisan and non-commercial effort that informs candidates and the public about the integrity of the content they consume. It works to assist candidates, campaigns, political parties, press organizations, fact-checkers, and social media platforms to identify, analyze, and report suspicious videos.
Potential deceptive media includes (but is not limited to) videos of candidates and other key figures making suggestive, damaging, or misrepresented versions of statements and actions. Reality Defender 2020 will not analyze intent; it will report manipulations to help responsible actors understand the authenticity of media before circulating misleading information.
“As a part of Reality Defender 2020, we hope to set a model for stakeholders in the upcoming election,” said the AI Foundation’s CEO and co-founder Lars Buttler. “The AI Foundation Nonprofit aims to discourage the generation and propagation of unsolicited and harmful synthetic content across technology companies, media outlets, political campaigns, and social media platforms. Our goal is to protect democracy, human agency, and truth.”
Reality Defender™ builds on the initial investment in the FaceForensics project at the Technical University of Munich (TUM) and brings together state of the art research at the AI Foundation, UC Berkeley, and other leaders in academia and industry on media forensics.
"The line between the real and fake is rapidly vanishing with enormous consequences to our society and democracy. It is critical that we develop and deploy technology to help us all regain trust in what we hear and see online," said Hany Farid, Professor, University of California, Berkeley.
Reality Defender’s intentional focus on diversity of research techniques and organizations comes from our aspiration to minimize bias and improve detection accuracies. Independent of Reality Defender, the FaceForensics project has attracted recent attention from Google and Facebook, furthering collaboration between technology companies for protecting information integrity.
Reality Defender 2020, available via rd2020.org, includes a submission page where videos can be entered for analysis by journalists and campaigns, and will also provide public information on known fakes and detection methods. Powered by Reality Defender™, RD2020 is a platform developed and maintained by the AI Foundation that combines detection methods from the AI Foundation and other partners.
For each submitted video, Reality Defender 2020 creates a report summarizing the findings of multiple forensics algorithms without speculating on the intent of the manipulations. An initial report is created within minutes of a video submission using automated analyses from AI models. This initial report is followed up with a comprehensive report including a manual review by expert forensic researchers and fact-checkers if the automated analysis detects manipulated content or if the models are not confident in their results.
About the AI Foundation
The AI Foundation, with support from many of the world’s leading top innovators, AI researchers scientists, engineers, and mission-driven investors, is a hybrid for-profit and non-profit with a mission to build a better future by giving everyone their own AI that shares their values and goals, while protecting them from harmful effects of AI technology.
The AI Foundation Nonprofit focuses on tools for detection and protection, while AI Foundation’s for-profit creates foundational technologies, products, and services to unlock the full human potential.
Reality Defender 2020 is supported by the AI Foundation’s Nonprofit Board of Directors, composed of key leaders across technology, media, entertainment, and communication sectors. Members include Twitter co-founder Biz Stone, venture capitalist and entrepreneur Cyan Banister, and academic leaders such as Prof. Rao Kambhampati and Prof. Hany Farid.