YouTube Patch Rundown: New AI Policies and Features

Updates on Monetization Audits and Content Policies

For Creators navigating the complexities of YouTube monetization, the completion of the September ESDA audits marks a significant milestone. These audits are critical for those dealing with restricted monetization on their content.

If you have older content that falls outside the window for automatic review, you are encouraged to take manual action. To resolve these restrictions, you can submit your content via discord.gg/YouTube, where it will be forwarded to Team YouTube for a formal re-review. This manual process ensures that legacy content is not unfairly penalized by outdated automated systems.

Looking ahead, YouTube is planning further relaxations of its content policies. Starting next month, the platform will focus on easing restrictions surrounding breastfeeding and non-sexually suggestive dancing. These changes suggest a move toward a more nuanced understanding of content that may have previously been flagged under broader, more restrictive guidelines, potentially opening up more monetization opportunities for Creators in these niches.

New Opportunities for YouTube Streamers

There is promising news for the streaming community regarding the YouTube Partner Program (YPP). YouTube is preparing to introduce an entry model specifically for streamers, similar to the initiatives seen last year.

While a precise date has not been set, the current expectation is that this will roll out in Q1 2024. The delay is strategic; YouTube is currently prioritizing the completion of reviews for small Creators under the “500-rule.” By ensuring the team is not overwhelmed by a simultaneous influx of applications, YouTube aims to provide a smoother onboarding experience for streamers once the current backlog is cleared.

On a more immediate note, some Creators have experienced technical issues with YouTube Live in recent days. Team YouTube has acknowledged these livestreaming problems and confirmed they are actively working on a resolution.

Clarification on Simplified Ad Controls

Recent discussions regarding “simplified ad control” have caused concern among smaller Creators, particularly those who are not yet monetizing. There has been significant worry that these changes might negatively impact watch time or other channel metrics.

However, after discussions with YouTube, it has been confirmed that these simplified ad controls apply exclusively to Creators who are already members of the YPP. If you are a non-monetizing Creator—meaning you are not in the partner program, even if YouTube is playing ads on your content—these changes do not apply to you. You can rest assured that your watch time and general channel standing will not be affected by this specific update.

The Rise of AI: Dreamtrack and Creative Risks

YouTube is aggressively integrating AI into its feature set, most notably with the introduction of “Dreamtrack” for Shorts. This tool allows Creators to generate original songs based on text inputs, utilizing the voice models of established artists.

While this represents a leap in creative capability, it introduces significant risks. The use of artist voice models for potentially “nonsense” content is a point of contention. For Creators using Dreamtrack, the most important takeaway is the necessity of strict adherence to community guidelines. Because AI-generated content can easily veer into prohibited territory, failing to follow these rules can lead to strikes. In the YouTube ecosystem, strikes are a serious matter, as repeated violations can lead to permanent channel termination.

AI Disclosure Obligations and Platform Integrity

Beyond specific features like Dreamtrack, YouTube is implementing overarching policy changes regarding AI-generated content. The primary focus here is transparency and the integrity of the viewing experience.

YouTube has announced new disclosure obligations and content labeling requirements. Creators who use generative AI to create content must clearly label it as such. This is not merely a suggestion; it is a policy requirement designed to prevent the audience from being misled.

The consequences for failing to comply with these disclosure obligations are severe. Sanctions may include:

• Demonetization of the specific content.
• Exclusion from the YouTube Partner Program.
• The forced removal of the AI-generated content.
• In extreme cases, the permanent termination of the channel.

The Challenge of Context in AI Moderation

Despite the severity of the penalties, the current policy formulations remain somewhat vague. The central challenge for both YouTube and Creators is the role of context.

To illustrate the difficulty of drawing a line, consider the following distinctions:

• Harmless vs. Problematic: Using AI to generate a joke about a president is generally seen as harmless. However, using AI to generate commentary or “facts” about sensitive geopolitical issues becomes problematic due to the risk of misinformation.
• Creative vs. Deceptive: Using Photoshop AI to enhance a vacation photo for a travel blog is a creative choice. Conversely, using those same tools to fabricate an image as evidence of a crime—even if it only appears for five seconds in a longer video—is a serious violation of trust and policy.

This creates a complex environment for Creators. There is currently a lack of clarity on where the “labeling” requirement begins. For example, if a Creator uses AI to optimize a thumbnail, refine a title, or polish a description, does that constitute “AI-generated content” that requires a label?

As these tools become integrated into the standard workflow—such as using ChatGPT to edit a script for better readability—the line between “AI-assisted” and “AI-generated” becomes blurred. Creators are advised to stay vigilant and keep a close eye on official updates from YouTube as these policies move from vague formulations to concrete enforcement.

Original transcript

Transcript

The September ESDA audits have been completed. So, if you still have content that’s too old for automatic review for restricted monetization, feel free to submit it to us via discord.gg/YouTube. We’ll then forward it to Team YouTube for re-review. Additionally, further relaxations are planned for next month, focusing on breastfeeding and non-sexually suggestive dancing. Early next year, there will also be an entry model for streamers in the YouTube Partner Program, similar to last year.

YouTube is actually waiting until all small creators are done with the 500-rule and all reviews are completed so the team isn’t overwhelmed. Then the streamers will be next, hopefully early next year. We don’t yet have exact numbers or a precise date, but Q1 2024 is expected. This is a promising development for YouTube streamers.

Also, regarding streaming: There have been some livestream issues on YouTube in recent days, and Team YouTube has said they’re addressing it. Good news—as soon as there are changes or updates, you’ll hear about it here.

Last week, we discussed how the simplified ad control might affect smaller creators. Now, I have good news: We discussed this with YouTube last week, and there was a bit of a sigh of relief. We were confirmed that there will be no change for non-monetizing creators. That means all the concerns we discussed last week are on hold for now. Because this simplified ad control only applies to creators in the YPP and not to non-monetizing creators who have ads played on their content. These changes do not apply to them. So, you won’t have any problems with watch time either.

Next, let’s turn to the topic of AI, as we’re starting right away with the AI function Dreamtrack for Shorts. With this, you can create your own song based on your text input using voice models from artists. There’s one way this can and will go wrong. That’s why I find it very questionable that well-known artists are lending their voices for this and essentially allowing it to be used for nonsense. Of course, YouTube has assured us that the guidelines will be followed. I would also advise you to always act within the guidelines, as acting outside of them can result in strikes, and strikes are never good because channel termination is permanent.

Regarding policy changes for AI content: YouTube announced in a blog post that policy changes will be introduced. These concern disclosure obligations and content labeling for creators using generated or AI-generated content. These changes are important to ensure the transparency and integrity of the platform. If the creator does not comply with the disclosure obligations, sanctions such as demonetization, exclusion from the partner program, removal of AI-generated content, and in the worst case, channel termination may follow.

However, the problem is that these policies are currently still very vaguely formulated. They have said that when we implement this, we will inform creators in good time. Felix Beuser mentioned a very interesting quote here: “Using AI to tell presidential jokes is fairly harmless. But letting them talk about geopolitical issues becomes problematic. Using Photoshop AI to create the perfect vacation picture for a blog might be interesting, but using the same tools to create an image as evidence of a crime used for only 5 seconds in an otherwise free video is again concerning.” I believe finding the right differentiation is a very difficult task. I hope YouTube will manage this fairly well.

Context plays a very crucial role. The question is, where do you draw the line? For example, where does it end when AI is used to edit thumbnails, titles, or descriptions? Is that already AI-generated content or something that doesn’t need to be labeled? This script here was also edited by ChatGPT to make it more understandable. It’s a very complex topic. Let’s discuss it in the comments, as I’d be interested to know your thoughts.

That’s all for today. Thank you for tuning in! I hope you took something away from it. See you next week for more YouTube updates. Until then, take care and keep going!