Are creator AI disclosures going to be a thing?

Hank Green has a new AI usage policy–and we’re wondering if other creators will follow suit.

For those unfamiliar, Green (who has made a number of videos about the rise of generative AI and its current and potential effects on humanity and the environment) recently admitted that he’d been using LLMs to such a degree that he was getting a “level of dopamine” that “is not healthy for me or good for the world.”

This admission followed scrutiny from viewers after Green used the phrase “and I appreciate the pushback” in a July 29 upload to his nonprofit educational channel Complexly.

Anyone familiar with LLM extrusions will know this phrase is a stock response from chatbots that often appears when prompters challenge the accuracy of generated information. The fact that Green said this mid-video, while he was alone in a room with no one giving “pushback,” made viewers suspicious that the script he was reading from had been generated.

Green responded to viewers’ concerns on X, saying he’d produced the video “under a ton of pressure” and had used OpenAI’s ChatGPT “for research on this script.”

In a separate post on Reddit, he went more in-depth, saying, “I have been relying too heavily on AI as a research aid. It can be very useful for this task, giving me access to a lot of papers I didn’t know existed really fast, but I think that has been to the detriment of my work because it has not given me the freedom to find all of my own ways into and around a topic.”

He said that after discussions with his wife Katherine and his brother/fellow YouTuber John, he’d decided to pause production on his personal channel, since part of the reason for his increased AI usage was the feeling that he needs to be “doing more and more and more and more.”

“It is very likely that you (the reader) and I disagree on how or even whether AI should be used. I think it makes sense for that disagreement to exist. But I have also always believed that my audience is more likely to be right than I am. You know what this feels like in a way I can’t. And I am so sorry for diluting myself in this way and harming the thing that we have all built together,” he said.

His personal channel is still paused, but Green used his Aug. 7 spot on his and John’s shared channel vlogbrothers to unveil his new personal AI policy.

That policy is:

1. No portion of any script will be written, edited, or outlined by an LLM.
2. The thesis of a video will always originate with a human.
3. No image or music in a video will be generated by AI. If something is accidentally included, best efforts will be made to remove it.
4. LLM outputs will never be trusted as a source.
5. Informational / educational videos will contain links to primary sources in the description.

In the video’s comments, the policy seems largely well received, although people do note that nowhere in here are the words “I will stop using AI.” These rules leave room for Green to continue using LLMs in production, so long as he thinks of his own video ideas and writes the scripts.

What we’re wondering is whether more creators will adopt a similar level of transparency and admit they use LLMs in various parts of production.

At the highest level, people who advise YouTubers are telling them to implement AI into their workflows, offloading as much as possible to achieve the very same “more and more and more and more” video output Green mentioned was detrimental to him.

Scroll YouTube’s homepage for 30 seconds and you’ll see at least a handful of videos with AI-generated thumbnails. Other videos on the platform were undoubtedly made with one of several widely available tools that scrape a YouTuber’s entire channel and generate new video ideas for them. Yet more creators are using LLMs to write their scripts, build their websites, and code their apps. Then there’s YouTube’s whole ‘deepfake yourself so you don’t have to bother being in your own videos’ thing.

Are all of these usages disclosed? Unless a creator is sponsored by an AI company, probably not. Tools from YouTube and other platforms do auto-label some generated content, like video “footage,” but thumbnails, scripts, and behind-the-scenes research are outside of that scope.

Creators who aren’t disclosing because they’re afraid of backlash might want to rethink their usage and trim down to the level stated in Green’s policy. Or they might want to write their own policies that explain their usage to viewers, and clarify which parts of their production are still human.

Looking at the Green’s policy, it’s clear he has learned from this experience, and wants to regain trust from his viewer base. If and how he accomplishes this over the next few months could give fellow creators a path to follow.