INSIDER

A 22-year-old TikToker fooled millions by pretending to be an AI-generated image - and he says it 'shows how little people understand AI and deepfakes in general'

INSIDER logo INSIDER 12.09.2022 21:06:08 mleighton@insider.com (Mara Leighton)
TikToker Curt Skelton, a 22-year-old college student, jokingly pretended to be an AI character. Many believed him. TikTok;@curtskelton

If you've scrolled TikTok recently, odds are good you've run into Curt Skelton. Whether you believe he's a real person, though, may be up for debate.

Skelton, who goes by @curtskelton on the app, went mega-viral for an August 25 video in which he claimed to be a sophisticated artificial intelligence deepfake - orchestrated by fellow visual effects creator Zahra Hussain (@zarooza) - rather than a flesh-and-blood human. 

The Tiktok, which has since been viewed 17.7 million times, features Hussain telling viewers that she used AI programs (some that exist, some that don't) to create the "fake character" - ending the video with an open-ended question for viewers: should she continue to post to the @curt.skelton account, or end the "experiment"? 

Many viewers assumed Skelton really was AI. It may not have helped that Hussain said Skelton wasn't made to look "hot" because she wanted people to "follow him for his jokes and visual effects." 

Some TikTokers responded playfully - seemingly in on the joke - while others expressed amazement or uncertainty: "I am having a crisis. Is she being honest or not," one popular comment read. 

But Skelton, a 22-year-old film student who currently works at a liquor store, is indeed real. Intended to be a joke playing off the fear that AI will one day replace digital artists, he told Insider he thought the video would be a "funny satire" - having AI literally replace him, a digital artist.

Skelton says he's made CGI jokes in the past - say, telling people his red hair turned brown as a child and that he's been editing it to appear orange ever since - and about half the viewers didn't get the joke then, either. So, he wasn't surprised when his AI prank fooled people, though he was shocked by just how many.

"I honestly was expecting maybe 20,000 views at most," he says, "So 50% of 20,000 not getting a joke isn't that bad or impactful. But 50% of 16 million on TikTok alone...Yeah, that wasn't the plan, lol."

Skelton, a student at California State University, Northridge, said in an August 29 video he was surprised more people didn't try to fact-check his existence by going to his other social media accounts. 

However, social media isn't necessarily conducive to consistent fact-checking. Viewers enjoying a one-minute video, quickly succeeded by a cache of other short videos, may not be motivated or capable of fact-checking each of the claims made in every single one they consume. (In 2021, teens and kids watched an average of 91 minutes of TikTok per day). Researchers have suggested that we may fall for misinformation like fake news thanks to 'lazy thinking' - or, lack of reasoning - more than bias.

Plus, TikTok's algorithm departs from competitors such as Instagram or Snapchat in its preference for novelty rather than familiarity. Instead of prioritizing accounts a user follows, TikTok's mysterious FYP routinely serves up content from creators a user doesn't follow - meaning they may have relatively less context on average. Some of the 17 million views were likely people new to Skelton's content (he told viewers he gained 100,000 followers from the viral video, Insider was unable to independently verify this claim).

Now that millions of people have watched Skelton's video - many of whom seemed to actually believe he was AI - he says critics have accused him of "maliciously spreading false information or trolling."

"After I realized I had fooled way more people than anyone should have, I realized the power I had. But with great power comes great responsibility," Skelton, who said he recently picked up a sponsorship from visual-effects software company Foundry, told Insider. "So I will be a LOT more careful in the future." 

"I think the fact that so many people fell for it shows how little people understand AI and deepfakes in general," Skelton continued, adding that he might use a more sarcastic tone or more obvious jokes in future videos. 

Representatives for Foundry did not immediately respond to Insider's request for comment.

mardi 13 septembre 2022 00:06:08 Categories: INSIDER

ShareButton
ShareButton
ShareButton
  • RSS

Suomi sisu kantaa
NorpaNet Beta 1.1.0.18818 - Firebird 5.0 LI-V6.3.2.1497

TetraSys Oy.

TetraSys Oy.