AI Junk Is Coming for Your YouTube Feed
AI junk has taken over Facebook, and it's coming to YouTube next. And, like a homeowner who doesn't seem to mind that the stranger at their door lacks a reflection, Google is letting it in.
At the company's Made on YouTube event in New York today, Google announced AI-powered features for creators, including AI-generated video titles, ideas, comment replies, and even full shorts. There are some non-AI additions, too, like a hype" system that works similarly to a "super like" and is intended to help small creators grow, but the bulk of the updates focus on ways to help creators shunt some of their burden onto robots.
Most notable is Veo, a new video generation model coming to Google's existing Dream Screen feature. While Dream Screen previously just used AI to generate backgrounds for YouTube Shorts, Veo updates the feature to the point where it can make full videos. In addition to being able to generate whole videos now, Google also promises Dream Screen will make more realistic backgrounds."
The catch is that these Veo-generated Shorts can only be six seconds long (for now, at least), so it'll be a while until AI starts making multi-hour essays. Still, it's worth being vigilant about what might and might not be real as the platform starts to officially embrace these features. Luckily, YouTube says Veo shorts will be watermarked.
Those hoping AI won't affect their more long-form viewing can't breathe easy just yet, though. Through a new Inspiration" tab in the YouTube Studio app, creators will now be able to view AI-generated video concepts, complete with suggested titles, AI-generated thumbnails, a script outline and a few example lines. It can be hard to keep up with changes to the site's algorithm, especially when you're pushed to upload longer videos more frequently-the Inspiration tab could be a way for Google to help creators know what is likely to rank highly, although it doesn't seem like video made with its help will be watermarked in any way.
AI will also help YouTubers once a video is out, integrating with another new feature-Communities-to help them chat directly with their audiences. While Communities as a whole operates more like a forum and creators will be able to reply manually, AI will also offer creators spotlights that will quickly catch them up on what's going on in their hub, plus give them suggested replies that YouTube says will be a starting point" when discussing with viewers. Again, AI-generated replies don't seem to have any markers on them.
Both of these features, while time-saving, present potential risks to the "You" in the "Tube," especially when it comes to the authenticity that so many YouTubers sell themselves on. Beyond being worried that replies from your favorite creator might have just been drafted by a robot, there's also the possibility that as the algorithm uses AI to suggest video concepts to creators, it will end up simply training itself once those videos do well, creating a sort of feedback loop of robots teaching robots, and limiting what types of videos creators can make while still getting enough views to stay afloat.
There are accessibility benefits to YouTube's new AI integrations. The company also showed off a new auto dubbing feature, which creators can turn on to have AI translate and dub their videos in a foreign language. This will help them reach a wider audience, although it's unclear what the quality of the translation and voice work will be like. If it's anything like YouTube's automatic captions, it might not work for every topic. Still, Google's Gemini Live voices do a decent job of sounding natural.
While more people are reporting that they've gotten access to some user-side YouTube AI, today's announcements were all about creators. That means it's going to be up to viewers to be more discerning about how much human involvement is in their content. The days of laughing at grandma for getting fooled by obvious AI art are in the past-zoomers and gen alpha need to be ready for it to come to their platforms of choice, and more subtly. While not everyone is against AI content as a rule, because of the speed with which it can be created, there's still a risk in inviting it to any social media site, in that it can quickly overwhelm all other types of posts. Ignoring it could quickly lead to a useless home feed.
YouTube, of course, isn't unaware of the risks of letting AI enter its feeds, and parent company Google is developing tools to make it more transparent when AI has been used. These new AI features will roll out slowly starting later this year or early next year, presumably after a lot of testing, and perhaps after the site has hung up a few garlic wreaths, just to be safe.