{"id":90523,"date":"2025-08-23T17:56:03","date_gmt":"2025-08-23T17:56:03","guid":{"rendered":"https:\/\/www.newsbeep.com\/au\/90523\/"},"modified":"2025-08-23T17:56:03","modified_gmt":"2025-08-23T17:56:03","slug":"youtubes-sneaky-ai-experiment-the-atlantic","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/au\/90523\/","title":{"rendered":"YouTube\u2019s Sneaky AI \u2018Experiment\u2019 – The Atlantic"},"content":{"rendered":"

Something strange has been happening on YouTube over the past few weeks. After being uploaded, some videos have been subtly augmented, their appearance changing without their creators doing anything. Viewers have noticed<\/a> \u201cextra punchy shadows,\u201d \u201cweirdly sharp edges,\u201d and a smoothed-out look to footage that makes it look \u201clike plastic.\u201d Many people have come to the same conclusion: YouTube is using AI to tweak videos on its platform, without creators\u2019 knowledge.<\/p>\n

A multimedia artist going by the name Mr. Bravo, whose YouTube videos<\/a> feature \u201can authentic 80s aesthetic\u201d achieved by running his videos through a VCR, wrote on Reddit<\/a> that his videos look \u201ccompletely different to what was originally uploaded.\u201d \u201cA big part of the videos charm is the VHS look and the grainy, washed out video quality,\u201d he wrote. YouTube\u2019s filter obscured this labor-intensive quality: \u201cIt is ridiculous that YouTube can add features like this that completely change the content,\u201d he wrote. Another YouTuber, Rhett Shull, posted a video<\/a> last week about what was happening to his video shorts, and those of his friend Rick Beato. Both run wildly popular music channels, with more than 700,000 and 5 million subscribers, respectively. In his video, Shull says he believes that \u201cAI upscaling\u201d is being used\u2014a process that increases an image\u2019s resolution and detail\u2014and is concerned about what it could signal to his audience. \u201cI think it\u2019s gonna lead people to think that I am using AI to create my videos. Or that it\u2019s been deepfaked. Or that I\u2019m cutting corners somehow,\u201d he said. \u201cIt will inevitably erode viewers\u2019 trust in my content.\u201d<\/p>\n

Fakery is a widespread concern in the AI era, when media can be generated, enhanced, or modified with little effort. The same pixel-filled rectangle could contain the work of someone who spent time and energy and had the courage to perform publicly, or of someone who sits in bed typing prompts and splicing clips in order to make a few bucks. Viewers who don\u2019t want to be fooled by the latter must now be alert to the subtlest signs of AI modification. For creators who want to differentiate themselves from the new synthetic content, YouTube seems interested in making the job harder.<\/p>\n

Read: ChatGPT turned into a Studio Ghibli machine. How is that legal?<\/a><\/p>\n

When I asked Google, YouTube\u2019s parent company, about what\u2019s happening to these videos, the spokesperson Allison Toh wrote, \u201cWe\u2019re running an experiment on select YouTube Shorts that uses image enhancement technology to sharpen content. These enhancements are not done with generative AI.\u201d But this is a tricky statement: \u201cGenerative AI\u201d has no strict technical definition, and \u201cimage enhancement technology\u201d could be anything. I asked for more detail about which technologies are being employed, and to what end. Toh said YouTube is \u201cusing traditional machine learning to unblur, denoise, and improve clarity in videos,\u201d she told me. (It\u2019s unknown whether the modified videos are being shown to all users or just some; tech companies will sometimes run limited tests of new features.)<\/p>\n

Toh\u2019s description sounds remarkably similar to the process undertaken when generative-AI programs create entirely new videos. These programs typically use a diffusion model: a machine-learning program that is trained to refine an extremely noisy image into one that\u2019s clear, with sharp edges and smooth textures. An AI upscaler can use the same diffusion process to \u201cimprove\u201d an existing image, rather than to create a new one. The similarity of the underlying process might explain why the visual signature of diffusion-based AI is recognizable in these YouTubers\u2019 videos.<\/p>\n

While running this experiment, YouTube has also been encouraging people to create and post AI-generated short videos using a recently launched<\/a> suite of tools that allow users to animate still photos and add effects \u201clike swimming underwater, twinning with a lookalike sibling, and more.\u201d YouTube didn\u2019t tell me what motivated its experiment, but some people suspect that it has to do with creating a more uniform aesthetic across the platform. As one YouTube commenter wrote<\/a>: \u201cThey\u2019re training us, the audience, to get used to the AI look and eventually view it as normal.\u201d<\/p>\n

Google isn\u2019t the only company rushing to mix AI-generated content into its platforms. Meta encourages users to create and publish their own AI chatbots on Facebook and Instagram using the company\u2019s \u201cAI Studio<\/a>\u201d tool. Last December, Meta\u2019s vice president of product for generative AI told<\/a> the Financial Times that \u201cwe expect these AIs to actually, over time, exist on our platforms, kind of in the same way that [human] accounts do.\u201d<\/p>\n

Read: What we discovered on \u201cdeep YouTube\u201d<\/a><\/p>\n

In a slightly less creepy vein, Snapchat provides tools for users \u201cto generate novel images<\/a>\u201d of themselves based on selfies they\u2019ve taken. And last year, TikTok introduced Symphony Creative Studio<\/a>, which generates videos and includes a \u201cYour Daily Video Generations\u201d feature that suggests new videos automatically each day.<\/p>\n

This is an odd turn for \u201csocial\u201d media to take. Platforms that are supposedly based on the idea of connecting people with one another, or at least sharing experiences and performances\u2014YouTube\u2019s slogan until 2013 was \u201cBroadcast Yourself\u201d\u2014now seem focused on getting us to consume impersonal, algorithmic gruel. Shull said that the modification of his videos erodes his trust in YouTube, and how could it not? The platform\u2019s priorities have clearly shifted away from creators such as Shull, whose combined work is a major reason YouTube has become the juggernaut it is today.<\/p>\n