Facebook’s new policy against manipulated media, announced January 8, 2020, didn’t spring from nowhere. It landed in the middle of a growing storm over deepfakes — videos and audio so convincingly faked that seeing is no longer believing. The platform’s move to ban deceptive edited content targets two specific kinds of trickery, and the definitions matter.
The first category catches media altered in ways an average person wouldn’t spot. This isn’t about obvious parody or clumsy edits.
It’s about subtle changes that make someone appear to say words they never spoke. Think of audio synced to mismatched lip movements, or a sentence spliced together from different recordings. The deception is the point, and the technology to do it keeps getting cheaper and better.
The second category goes after content built by artificial intelligence or machine learning that merges, replaces, or superimposes elements onto video to make it look authentic. That means face swaps, background changes, or altered mouth movements matched to a fake audio track.
These tools, once the province of Hollywood studios, now run on consumer-grade software. A person with a laptop and a few hours can produce a fake that fools a casual viewer. Facebook Global Policy Management President Monika Bickert laid out these categories as part of a broader enforcement framework.
The company stated it would ban deepfakes and misleading manipulated media outright. The policy is a direct response to fears that such content could be weaponized — to smear a politician, spread a false rumor, or incite panic. The 2020 U.S. presidential election was months away.
The timing was no accident. The problem Facebook faces is definitional.
What counts as deceptive? A satirical edit is usually obvious. A propaganda video that swaps a leader’s face onto someone else’s body is not.
The company’s policy tries to draw that line based on whether the average person would be fooled. That’s a subjective test, and it will be tested.
Enforcement is the hard part. Facebook has billions of users and uploads mountains of video every day. Catching every deepfake before it spreads is impossible.
The policy is a statement of intent, a promise to act when content is flagged or discovered. But the technology to create fakes is evolving faster than the tools to detect them. The policy may deter amateurs.
It may not stop state-backed disinformation campaigns with real budgets and technical talent. The company’s announcement on January 8, 2020, acknowledged that the problem is not going away.
Artificial intelligence tools that alter video, audio, and visual content are becoming more sophisticated. The policy bans content that makes it appear authentic while conveying falsehoods. That’s a high bar for enforcement, but it’s a necessary one.
Facebook’s move came amid broader pressure on social media platforms to police misinformation. Critics have long accused the company of moving too slowly.
The deepfake policy is a step, but it’s one step in a long race. The technology will keep advancing. The policy will need to keep up.
For now, the company has drawn a line. Whether it can hold that line is an open question.
{“@context”: “https://schema.org”, “@type”: “NewsArticle”, “headline”: “Facebook’s New Policy Targets Undetectable Deepfake Videos and Audio”, “description”: “Facebook’s new policy against manipulated media, announced January 8, 2020, didn’t spring from nowhere. It landed in the middle of a growing storm over deepfake”, “datePublished”: “2020-01-08T15:00:00+00:00”, “author”: {“@type”: “Person”, “name”: “InfoPulse Today”}, “publisher”: {“@type”: “Organization”, “name”: “InfoPulse Today”}}
