Hank Green found the AI problem that YouTube labels can’t catch



YouTube currently requires that content creators let viewers know “when they use AI to meaningfully alter or generate photorealistic content.”

The policy draws some strange boundaries. It applies to “AI-generated music” (not photorealistic) but not to “riding a unicorn through a fantastical world” (this could be photorealistic, though it is not plausible). YouTube then summarizes the policy in a different way: “Realistic AI content and meaningful changes require disclosure, while non-realistic or minor edits don’t.”

But AI uses that require no disclosure can include everything from “idea generation” up through “production assistance, like using generative AI tools to create or improve a video outline, script, thumbnail, title, or infographic.” Creators are free to clone their own voices for voiceovers. They can also use “AI-generated or altered animation of a missile in a fully animated video.”

This results in some odd scenarios. A thrilling 10-second video that shows me riding my AI-generated steed through the horse-killing-fart swamps of Soylentius IV? No disclosure, even though the entire thing is AI-generated. But when I add an AI-crafted lute ballad about the grave dangers I faced in those swamps? Mandatory disclosure.

The policy gap becomes more consequential when you imagine a 30-minute video attempting to sway people’s views on geopolitics. AI could generate the premise, do the research, and write the outline. My own AI-cloned voice model could read the AI-drafted script. I could even use AI-generated missile animations. Must I disclose the rampant AI use that drove this entire project? Apparently not.

But having an AI help in these ways imparts a certain feel and logic to projects, even if the final result is not fully “AI-generated.” Humans approaching topics without AI might find sources through quite different paths, and they might have to read and process more material to get there, giving them a different kind of understanding. They might also note very different things as important, thus creating different outlines of the same material. They might pepper a script with jokes or digressions not usually suggested by an AI. And they might read a script aloud in a more natural way.



Source link

  • Related Posts

    Meta Introduces Muse Code, Its Take On A Coding Agent

    The terminal-based coding tool is powered by a new AI model, Muse Spark 1.2. Meta has announced an early beta of Muse Code, a new coding agent meant to compete…

    Meta Is Challenging Claude Code and Codex With New Muse Code

    Meta is making strides to catch up to AI heavyweights like OpenAI and Anthropic. The company announced on Wednesday a new agentic coding tool and the new AI model that…

    Leave a Reply

    Your email address will not be published. Required fields are marked *

    You Missed

    On the road for the CFL – In This Corner

    Carney says tone with U.S. already ‘quite firm’ as trade talks continue

    Carney says tone with U.S. already ‘quite firm’ as trade talks continue

    Ontario byelections give Doug Ford a chance to prove the haters wrong

    Watch live: Trump speaks at Red Rock Resort in Las Vegas

    Watch live: Trump speaks at Red Rock Resort in Las Vegas

    Meta Introduces Muse Code, Its Take On A Coding Agent

    Meta Introduces Muse Code, Its Take On A Coding Agent

    The Real Reason Why No US Airline Ever Operated The Airbus A380

    The Real Reason Why No US Airline Ever Operated The Airbus A380