YouTube is testing AI to "improve" video quality

In recent years, artificial intelligence (AI) has become an integral part of digital platforms, including video hosting services. YouTube, the world’s largest video-sharing platform, has begun using AI to optimize and modify video content. Recently, some users noticed that YouTube Shorts looked different: sharper, smoother, or with altered visual effects. It turned out that YouTube was experimenting with AI-driven video enhancement — without informing creators or viewers.

What AI Does on YouTube

YouTube applies AI algorithms primarily to Shorts, aiming to make videos more visually appealing. These algorithms automatically:

• Increase image sharpness

• Smooth textures such as skin and background

• Adjust lighting and shadows

• Enhance contrast and color saturation

These changes occur without the creator’s involvement. The platform is testing how AI-driven improvements affect viewer engagement. Reportedly, YouTube is also experimenting with real-time quality adjustments, aligning videos with its own visual standards.

Moreover, YouTube is developing AI tools to assist in content creation. For instance, Veo 3 technology can generate short videos from static images or text descriptions. While this opens opportunities for easy content production, it also raises concerns about authenticity.

The Transparency Problem

A major issue is the lack of transparency. YouTube does not notify creators when their videos are altered by AI. Viewers themselves noticed unusual changes — overly smooth skin, exaggerated shadows, or visuals that did not match the original. YouTube later confirmed that these modifications were part of ongoing experiments.

The platform’s main goal is to test AI technologies under real conditions, studying their effect on watch time and engagement. However, the absence of notifications leaves creators without control over their content and viewers unsure if they are watching the original work.