Facebook observed that there have been photos that had been manipulated through Photoshop or use deepfakes to deceive people. The social media networking site aims to address the spread of deceitful media outputs on the platform. On 6 January 2020, Facebook announced it will combat deceitful media outputs such as deepfakes.
Their first solution is to partner with experts from around the world to tackle manipulated media. Through thorough consultation and discussion with experts, Facebook decided to remove manipulated media with 2 criteria:
However, the policy doesn’t cover any parody or satire related videos that were edited or had words omitted.
Any content that contains a deepfake aligned with other violations will be removed as part of Facebook’s policies. Any video of those kinds would still be subject to investigation by Facebook’s independent third party fact checkers. Once tagged as false, the circulation of the post will be limited and isn’t eligible to become an ad.
Facebook also mentioned that it removed AI-generated content from fake accounts that masked their identities. To combat deepfake content from spreading, Facebook held a challenge to users in helping them identify any deepfake. Additionally, Facebook collaborated with Reuters to help media identify deepfakes through a free online training course.