Reality Defender 2020

A FORCE AGAINST DEEPFAKES

A non-partisan and non-commercial effort to help reporters and campaigns uphold truth and ethical standards.

WHAT

Detect Synthetic Media With AI

1. UPLOAD

Upload a piece of media to be evaluated

2. THE PLATFORM ANALYZES

Reality defender runs a series of models to detect manipulation

3. REVIEW RESULT

Review each models probability of the media being manipulated

WHY

Fight Against Deepfakes

There is increasing evidence of synthetic media being abused in media, technology, and political campaigns. AI-generated deepfakes could become even more harmful in the 2020 election, used to create highly suggestive, damaging, false versions of candidates' statements and actions. It's important to have a concerted effort to establish truth.

HOW

We are doing this in a completely inclusive, non-partisan manner by bringing together the best minds in media forensic research, AI, technology and journalism. The RD 2020 tool is a realization of this motivation.

Global AI Collaboration

TECHNOLOGISTS | PLATFORMS | POLICY MAKERS | EDUCATORS

Reality Defender is powered by a global collaboration of stakeholders invested in the fight against misinformation.

Forensic researchers with advanced
AI algorithms and mutual
data sharing.

Leaders in technology, media, and
entertainment pledge, inform,
and collaborate.

DATASET

Largest Deepfake Detection Dataset

Released in 2017 in partnership with TUM, FaceForensics and FaceForensics++ are the first large scale video data sets containing faces that can be used to study image or video forgeries. With over 500,000 manipulated frames, FaceForensics was one of the first tools used by researchers to enable the public to participate in detection of deep fakes.

The AI Foundation Nonprofit's Mission

To anticipate and counteract the dangers of AI by giving each of us the tools to protect our lives, dignity, and freedom.

ENDORSEMENTS

“There is an urgent question, how do we prepare to co-exist with AI?

The AI Foundation recognizes the duality of these powerful new tools and sees clearly how we must forge ahead”

- Biz Stone

Co-founder, Twitter

Director, The AI Foundation Non-Profit

"As the sophistication of deepfakes grows, thankfully, our sensitivity to deepfakes also reduces - making the near term the most vulnerable - and thus actionable - compared to the long term.

This is why AI Foundation is dedicated to help fight this problem for the 2020 elections.”

- Rao Kambhampati

Chief AI Officer, The AI Foundation

Past President, Association for the Advancement of Artificial Intelligence

“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.”

- Hany Farid

Professor, University of California, Berkeley

Board of Directors, The AI Foundation Non-Profit

GET INVOLVED

Request Access for Journalists & Campaigns

Invite-only submission page where potential fake content is entered for analysis.

Utilizing independent AI-based and human-vetted analysis, Reality Defender 2020 scans and verifies whether content is fake, manipulated, or original. A report on each submitted video summarizes the findings without speculating on the intent of the manipulations.

Contribute to Our Mission

Help develop and scale our technology and research. We’re always looking for new detection models, new expertise, infrastructure and donations.