Free AI Video Generators Are Nearing a Crucial Tipping Point

Advances in artificial intelligence (AI) technology mean that generating convincing images and memes is now relatively easy. Similar developments in video are both underway and lagging behind image generation. Although generating individual frames is computationally expensive, a growing range of AI tools can simulate movement and produce convincing animation. These may herald a shift towards the widespread use of AI videos that raise questions about the authenticity of all connected media. At the same time, examples of how generative AI can repurpose harmful or degrading images and stereotypes highlight the potential for these videos to exacerbate already divisive social, political and cultural issues.

AI-generated images and videos can be created quickly and at a relatively low cost, as algorithms continue to improve. Image-generating programs can quickly generate realistic-looking graphics from text inputs, with various tools and knockoff generators now readily available. Reddit and other online forums offer various examples of AI images that are both compelling and surreal. These developments raise warning flags in terms of the authenticity of online media.

Although AI-generated video is still somewhat lower in quality, a range of tools used in the making of memes and similar short clips emphasises that the technology is advancing. For instance, Runway ML’s latest tools can efficiently apply stylistic effects to videos, while Luma AI uses neural radiance fields to create 3D scenery from 2D photographs. Nonetheless, problems arise in creating coherent narratives from individual frames, which remains computationally expensive and beyond the AI’s capabilities at present.

The rise of generative AI poses significant ethical and social risks, as manipulated images and videos can lead to division, confusion and propagate poor information. Ultimately, we will likely find ourselves questioning whether the visual media we receive is authentic and truthful. While video has not yet resulted in the kind of deepfakes that we have seen created with images, the pace of technological change suggests that this too may become more mainstream soon.



Source link