Creatives grapple with AI-assisted vs. AI-made labels

The distinction between AI-assisted and AI-made is becoming harder to define as using the technology becomes a routine part of making content.

Why it matters: Creators, platforms and consumers are drawing different lines around how much AI involvement is acceptable and when it should be disclosed.

What they're saying: YouTube CEO Neal Mohan said last month that the platform doesn't view AI-generated content as inherently low quality.

"YouTube's been around for 20 years, and our mission is to give everyone a voice, show them the world," Mohan said. "It's a place without gatekeepers."

"AI versus not is kind of an artificial sort of distinction," he added. "Every video may have AI in it, just like almost every piece of music has something synthetic in it today versus 100 years ago when it didn't."

Zoom in: Music platforms have started to formalize how much AI involvement warrants a label. YouTube Music announced this week it's rolling out labels for songs created with generative AI.

Music partners can classify tracks as fully AI generated, partly AI generated or not AI generated, though listeners won't see that distinction. An AI icon expands to say, "This audio was altered or fully generated using AI."

Under YouTube's guidelines, songs combining AI lyrics with human performances fall into the partly AI-generated category, while pitch correction like auto-tune doesn't count as generative AI use.

Zoom out: For other formats, platforms have been developing different rules.

YouTube made labels more prominent in May for "photorealistic and meaningfully AI altered or generated content." Routine production assistance, such as AI-generated scripts or captions, is exempt.

TikTok requires labels for realistic AI-generated images, audio and video, including significant alterations to someone's likeness or speech.

Meta requires disclosure of photorealistic AI-generated video and realistic synthetic audio. It moved labels for content modified or edited with AI into posts' menus after criticism that minor edits were being treated like fully AI-generated content.

State of play: Some creators are keeping AI behind the scenes. Dhar Mann told Axios at Cannes that his team experimented with AI-generated effects in videos but received audience backlash.

His studio currently avoids AI-generated effects in its finished videos, instead using AI agents to analyze audience data, optimize workflows and identify promising storylines.

Jared Gutstadt, founder of Rebel Audio, uses AI to automate editing, clipping, packaging and distribution.

"I need AI to do all the things around my work so that I can work," Gutstadt says.

Filmmaker Zack London is preparing to release "Gods Don't Give Gifts," a feature film with AI-generated visuals but human voice actors, musicians and a creative team of roughly 25 people.

London says it's inaccurate to describe the film as fully AI generated because humans make every creative decision under his direction. He considers it AI assisted.

"I don't think there's an agreed-upon definition, but I see it more as a spectrum," London says.

What to watch: Some platforms are giving users more control over how much AI-generated content they see.

TikTok lets users adjust how much AI-generated content appears in their For You feeds.

Meta allows users to filter or request less AI-generated content across their Facebook and Instagram feeds.

YouTube doesn't currently have such a setting. But Mohan hinted at a customization option when asked about AI filters at last month's Made on YouTube event. He pointed to custom feeds, a new feature that lets viewers ask for particular types of content.