Exploring the creative possibilities of generative AI on YouTube, here’s a preview of upcoming updates in the next few months and the new year.

Generative AI holds the potential to enhance creativity on YouTube, reshaping the experience for both viewers and creators. Yet, alongside these possibilities, it’s crucial to uphold our responsibility in safeguarding the YouTube community. All content, regardless of its origin, must adhere to our Community Guidelines. Recognizing that AI introduces new challenges, we are in the early stages of refining our strategy, ready to adapt as we learn more.

In the upcoming months and the new year, YouTube will implement the following measures:

Disclosure Requirements and New Content Labels: We recognize the importance of maintaining a healthy information ecosystem on YouTube. While our existing policies already prohibit misleading content, the potent storytelling capabilities of AI may introduce new risks. To address this, we will introduce disclosure requirements and new content labels. This aims to ensure transparency, particularly when AI is employed to create content that could potentially mislead viewers, especially if they are unaware of the synthetic nature of the video.

To address this concern, we’re implementing updates in the next few months to keep viewers informed when the content they’re watching is synthetic. Specifically, we will require creators to disclose if they’ve produced realistic altered or synthetic content, utilizing AI tools. When creators upload such content, they’ll have new options to indicate that it includes realistically altered or synthetic material. This disclosure is particularly crucial for content discussing sensitive topics like elections, ongoing conflicts, public health crises, or involving public officials.

Creators who consistently fail to disclose this information may face consequences such as content removal, suspension from the YouTube Partner Program, or other penalties. We will collaborate with creators beforehand to ensure they understand these new requirements.

To notify viewers, two methods will be employed:

1. Description Panel Label: A new label will be added to the description panel, indicating that some of the content has been altered or synthetic.
2. Video Player Label (for sensitive topics): For certain types of content, especially those addressing sensitive topics, a more prominent label will be applied directly on the video player. This aims to provide clear visibility and transparency regarding the nature of the content.

In certain cases, a label alone may not sufficiently mitigate the risk of harm. Therefore, synthetic media, even if labeled, will be removed from our platform if it violates our Community Guidelines. For instance, a synthetically created video featuring realistic violence may face removal if its intent is to shock or disgust viewers.

As we implement these updates, content generated by YouTube’s generative AI products and features will be distinctly labeled as altered or synthetic. This step aims to enhance transparency, ensuring that viewers are aware when content is created using AI technology on our platform.

New Options for Creators, Viewers, and Artists

We value the ongoing input from our community, including creators, viewers, and artists, regarding the potential impact of emerging technologies. This is especially pertinent when it comes to instances where someone’s face or voice can be digitally generated without consent or used to misrepresent viewpoints.

In the upcoming months, we’ll enable the ability to request the removal of AI-generated or other synthetic/altered content that simulates an identifiable individual, including their face or voice, through our privacy request process. Content removal won’t be universal, and we’ll assess various factors, such as whether the content is satire, if the requester can be uniquely identified, or if it involves a public figure.

For music partners, we’re introducing the option to request the removal of AI-generated music content mimicking an artist’s singing or rapping voice. Considerations for removal requests will include factors like whether the content is subject to news reporting or serves as analysis or critique of synthetic vocals.

Deploying AI Technology for Content Moderation

YouTube employs a combination of human reviewers and AI technology, with AI classifiers aiding in the detection of potentially violative content at scale. The integration of generative AI has proven valuable in identifying novel forms of abuse, enhancing the speed and accuracy of content moderation systems. This helps reduce harmful content exposure for human reviewers.

Building Responsibility into AI Tools and Features

Our approach to developing new AI tools prioritizes responsibility over speed. We’re committed to refining technology capable of moderating content at scale and establishing guardrails to prevent AI tools from generating inappropriate content. Recognizing the inevitable attempts by bad actors to bypass these measures, we remain dedicated to incorporating user feedback and continuous improvement.

The AI Transformation is Upon Us

As we embark on the journey to unlock innovation and creativity with generative AI on YouTube, we are mindful of striking a balance between these benefits and the safety of our community. This transformative moment requires collaboration with creators, artists, and the creative industries to shape a future that benefits everyone.