YouTube continues to make changes to the user experience. Some of these features intentionally make YouTube viewing worse, but others, like the new custom YouTube feed feature, have proven to be really helpful. But one area where YouTube is particularly struggling is the influx of AI content. Thanks to new features like Reimagine, which allows you to shoot and recreate short videos using AI prompts, and the influx of AI slop to the platform in general, browsing YouTube has become a minefield for those looking for authentic human-driven content. But YouTube wants to fix that.
Starting in May, YouTube will add more prominent labels to videos featuring photorealistic AI. Long-form videos that currently have a label in their description will now have a label right below the video player, making it easier to review. YouTube Shorts videos, on the other hand, will now be labeled directly on the video itself. YouTube says the changes are aimed at giving viewers more context at a glance.
Additionally, we are also launching a new feature that automatically detects signs of AI in videos. If they are found, the system will automatically add the AI label. Many people commented on YouTube’s announcement video, thanking the company for increasing its transparency, but others argued that adding labels alone wasn’t enough. Creators can still manually mark their videos as using AI. This is an option starting in 2024.
YouTube users want platform to take tough stance on AI
One user even wrote that the platform should allow you to exclude AI videos entirely, especially as a subscriber to YouTube Premium, which had its price hiked earlier this year. Others agreed that it should be possible to add a filter to search results to completely hide AI videos. This makes it less likely that users will actually click on the AI video. One thing users seemed to particularly agree on was that YouTube shouldn’t stop at photorealistic videos. Also, anything that uses AI narration, AI scripts, or even AI editing should be labeled.
Of course, proving all of this is another matter entirely, but users seem to think YouTube can pull it off. One user also mentioned that adding fact-checking information would be a useful feature, as it could help reduce the random AI “fact channel” that uses AI for everything.
At this time, YouTube does not specify whether labels should be clickable for users to learn more. But that would be natural for them. YouTube said creators who are caught by its automated system and incorrectly identified as using AI can appeal. However, videos whose C2PA metadata primarily indicates a source that is considered AI, or videos with watermarks, such as videos created with Google Veo, cannot be claimed.
