DID PRESIDENT TRUMP share a fake clip of Nancy Pelosi? The seemingly simple question is a vexing one to answer. Passing judgment on his behavior is less challenging.
Conservative accounts on social media circulated a clip this week deliberately distorted to make it seem as if the speaker of the House was slurring her speech: “Drunk as a skunk,” commentators declared. The video, which some declared a “deepfake,” employed much too simple technology to merit that term. Deepfakes use artificial intelligence to synthesize human images into a reality that is entirely fabricated; the smear of Ms. Pelosi merely slowed down parts of an existing interview and modified her pitch.
The clip Mr. Trump tweeted alongside the words “PELOSI STAMMERS THROUGH NEWS CONFERENCE” was part of the same narrative, but it was not distorted, or even doctored, so much as it was edited. The clip splices together short segments of Ms. Pelosi (D-Calif.) stuttering in a lowlight reel that offered a misleading impression of a perfectly coherent 21-minute news conference. Mr. Trump did not make this video, or pull it from the right-wing fever swamps of social media. He took it instead from the fever swamp of Fox Business Network.
The clamor for firms such as YouTube, Facebook and Twitter to remove or limit the distribution of these clips as misinformation invites a vexing debate about what counts as fake in the first place. The “slurring” video, accompanied by manufactured accusations of drunkenness, may fall on one side of the line. The stammering video may fall on the other. But drawing that line at all has far-reaching implications. People edit videos all the time, sometimes for fun and sometimes to prove a political point. When does editing become doctoring, and when does doctoring become distorting? Is distorting always impermissible, or does it depend on intent, effect or something else altogether?
These difficulties both are caused by and contribute to the erosion of trust in today’s America, where it is hard to say what there is more of: false cries of “fake news,” or viral “news” that is actually fake. Technology certainly has helped this issue along, providing both an easy means to craft propaganda and an easy means to promote it. The increasing sophistication of image editing that creates the threat of actual deepfakes filling the Web will make that worse.
In the best of circumstances, the emergence of these tools for mass deception would be disturbing. It becomes absolutely alarming at a time when America is led by somebody who is intent on deceiving. The role of a responsible leader is to be a bulwark against an assault on truth, yet instead Mr. Trump is a battering ram. That’s not a problem Twitter or any other platform can solve.