Transcription
Right. Then YouTube's CEO, hi Neil, has revealed how the platform plans to handle AI generated content, including a brand new way to monetize that will be exclusively available to AI content creators. But how on earth do we get there?
>> And I think that's sort of like a a priory distinction between, well, it was AI generated, therefore it must be violative versus not actually is not the way to do it.
>> That's right. No bans, no purges, no Skynet shutdown sequence. Instead, they're building policies and technical frameworks to support it. Yep, you heard that right. YouTube is actually trying to make AI content legit. And to some of you, that might sound disturbing. Let us know what you think in the comments below. But trust me when I say there is no turning back. You can hate it with a passion. You can choose not to use it. You can even bury your head in the sand and forget about it, but it's not going anywhere and we're all going to have to accept that. So, let me explain very broadly how this is going to work and what it means for creators like you. Because this is not just policy talk. This is changing the rules of the creative game.
YouTube CEO Neil Mohamm was speaking recently at the All-In Summit and he just dropped the platform's official stance on AI and his main idea is this.
>> With the growth of these AI enablement tools, you're going to see that like the distinction between like was it completely AI generated, was it AI assisted, etc. is going to be a continuum.
>> AI content isn't one thing. It's not either good nor evil, fake or real, Skynet or Pixar. It's a spectrum. Some creators use AI as a tool to help edit videos, write scripts, or even design thumbnails, like many of the tools we have here at VidIQ. Exclusive viewer discount available in the description below. Others, on the other hand, go full AI mode, generating entire videos with synthetic voices, faces, and personalities. And the line between the two already blurred and getting fuzzier by the day. So YouTube's response isn't to draw a line in the sand. It's to accept the reality and build guard rails around it.
>> You know, YouTube has these, you know, community guidelines. We try to be transparent about them. We publish them on our website. And it's really just about giving as much transparency to the users as possible.
>> What Neil's basically saying is AI is fine within the well-established rules of YouTube. What's changed is that it's easier and quicker than ever to challenge those rules through AI. So, how do you build a platform around this blurry content? Step one, transparency.
>> One of the things that we do there is a label that says that you know, you know, # AI generated and um we will put that sort of we call it sort of front of the box. We will literally put that on the video and in some cases usually it's in the metadata. Um that's obviously not foolproof.
YouTube now requires creators to label AI generated content. It's not a warning. It's not a scarlet letter. It's just a heads up for viewers, an advisory, a yellow road sign. Now, as Mohan admitted there, this won't catch everything and it's not going to when as of time of recording, the AI generated content label seems to be voluntary. I can watch pretty much any video about Sora 2 on YouTube right now and it won't have a label in the description. And here's a question for you. YouTube said that they might put a AI label on the video itself. I have yet to see this. Have you? So maybe so far YouTube gets a B minus on the implementation of AI content transparency. But the important thing to note is YouTube want to label it, not bury it. Let the audience decide. YouTube state that AI generated content will not be automatically punished. Let me repeat that for all of the inevitable Reddit threats. AI content is not inherently a community guideline violation. Repetitious and reused content is, but that's not exclusive to AI generated content. And it clearly states in these support pages about disclosing content as AI generated that it will not impact the video's audience or its eligibility to earn money. To put it into its simplest terms, YouTube cares what you publish, not how you made it, unless it's blatant copyright theft, which is a topic for another video.
Hey Neil, when it comes to AI and the concerns around it, YouTube are a lot more focused on step two, protection of identity. Whether you're a Hollywood celebrity or a YouTuber or or an artist, the the thing that I hear over and over that they really care about is their likeness. This is the real problem YouTube is tackling. Not the tools themselves, not the tech, the theft, which Sora 2 has definitely started to flex and bend around.
>> It's been 30 years since I passed. Let me rest, please.
>> So, how does YouTube solve a problem of someone with malicious intent taking your face, your voice, your audience, your reputation, and doing whatever they like with it.
>> So, Content ID is, think about that metaphor as it applies to AI.
>> In case you don't know what Content ID is, once your channel is in the YouTube partner program, YouTube will automatically detect when other channels have used your videos on their channel. This portal allows you to take action should you want to.
>> We are working on this notion of I call it sort of likeness detection where if it's a chimoth face um the algorithm should be good enough to actually detect that and then give you a choice as to whether that should come down.
>> I've barely seen any screenshots of this tool yet and these were the best ones I could find. one of the portal screen here and one of the request removal screen and it does look pretty much like the content ID system. Brilliant if it works. And when does it come out? Because Sora 2 is running a mock right now and we here at VidIQ don't have this likeness detection tool yet. Let us know if you have it in the comments below.
But here's the power move that I didn't see coming. Creators might be presented with a choice when their likeness is detected on the YouTube platform. Option one, of course, is take it down. Simple as that. But what about this alternative?
>> Or, you know, there might be some creators that choose to monetize and create ownership of it.
>> You can take ownership of it.
>> So, that's right. You could literally earn money from someone else's AI remix of your voice and face. Now, imagine that scaled up across music, podcasts, educational videos, you name it. Does this turn AI from a threat into a business model? Possibly. Will it be perfect? No. Will it be abused? Absolutely. But if it's impossible to stop AI and it doesn't make sense to ban it, isn't the next best option to allow creators to benefit from it?
You open up the YouTube app, you hit that plus button, um you can type in a text prompt and it will generate a video for you using our VO models. Right? So that that is already happening.
>> Let's zoom out, shall we? That plus button is just the start of what YouTube calls unlocking opportunities, unlocking efficiencies, empowering human creativity. And now we have the labeling system, likeness detection. And what do I call this? AI licensing. These aren't just policy updates. It's infrastructure. And infrastructure is how platforms shape the future. For YouTube, it's not about fighting AI. It's about legitimizing it, policing it from a distance, providing it to creators, and leveraging it. And if you're hearing all of this for what feels like the 20th time, shaking your head and saying to yourself, "Nah, that ain't it, Chief." Well, I'm afraid it is because he's a chief and he's just told you what YouTube is going to do. So, what are you going to do?