TikTok creator program incentivizes AI misinfomation
Published Date: 3/26/2024
Source: axios.com

TikTok's cash-for-creators program could speed the spread of spammy misinformation on the platform as video makers lean on AI to hook viewers with outlandish images, conspiracy theories and hokum.

Why it matters: The crazier the video, the higher the engagement — and now, more engagement directly translates to more dollars.


The big picture: While U.S. lawmakers continue to worry that China is weaponizing TikTok to spread propaganda, AI tools are helping people spread a wide blanket of untrustworthy information on both TikTok and other social media as a crucial election year gets underway.

Driving the news: TikTok's new Creator Rewards Program, which came out of beta on March 18, pays creators "based on qualified views" and RPM, which is the average gross revenue per 1,000 "qualified" views, according to the company.

  • In addition to increased earnings for "originality," TikTok says the new program uses an "optimized rewards formula" to pay more for videos that are longer and optimized for "search value" and "audience engagement."

The big picture: Building engagement, optimizing for search and producing more, longer content are all tasks that generative AI can streamline.

  • Generative AI is great at making stuff up. If you're not worried about historical inaccuracy, hallucinations, bias or copyright law, the new AI tools are perfect for helping creators make videos quickly.
  • While Google, Meta, OpenAI and Microsoft have dedicated red teams to identify and address safety issues with their generative AI tools, anyone can easily download free tools that don't aim for such safeguards.

How it works: AI tools can automate the entire process of making videos, writes Jason Koebler in 404 Media.

  • ElevenLabs can turn scripts written by ChatGPT into voiceover speech so creators don't need to be performers or even use their own voices or likenesses in videos.
  • CapCut uses AI to edit videos and create transitions. The app is owned by TikTok parent ByteDance and is currently the seventh most-downloaded in the U.S.

To join the new creator program you must be at least 18 years old, have at least 10,000 followers and have at least 100,000 video views in the previous 30 days.

  • You must also have a U.S. account — but there are now sites that will sell you one.

Between the lines: According to misinformation researcher Abbie Richards, writing in Media Matters, the incentives from TikTok's beta creator program combined with weak content moderation rules mean the platform is already "flooded" with conspiracy videos made with free AI tools.

  • Richards found conspiracy videos featuring vampires, wendigos and King Kong.
  • A simple search for "conspiracy theories videos 2024" on TikTok yields videos that feature AI-generated images of Trump and Biden or cite AI-generated Joe Rogan footage to back up claims.
  • Koebler and Richards both found YouTube tutorials on how to use the tools as well as pyramid scheme-like communities of creators selling access to tools and templates with promises to make thousands of dollars a month with their methods.
  • Some of the videos contain labels or hashtags noting that they are fake or AI-generated, but some do not.

Yes, but: "Conspiracy theories are not eligible to earn money or be recommended in For You feeds on TikTok," a TikTok spokesperson told Axios.

  • "Harmful misinformation is prohibited and we apply firm account standards to our rewards program to encourage high-quality, original content," the spokesperson added.
  • TikTok videos often straddle the line between information and parody, making it hard to distinguish silly gags from potentially harmful lies.
  • TikTok's community guidelines forbid "content that tricks or manipulates others as a way to increase engagement metrics."

The problem of AI misinformation isn't confined to TikTok. A recent study from the Stanford Internet Observatory found that Facebook page operators are using AI-generated images to go viral and then push visitors towards spam content.

  • The study also found that Facebook's content algorithm is pushing this content into users' feeds.

By the numbers: The proliferation of misinformation on TikTok isn't new.