Famous actor Will Smith shared a parody of the pasta eating video, which went viral last year and was created with artificial intelligence by a Reddit user, on his Instagram account. “This is out of hand!” The video, published with the title, shows the progress of artificial intelligence in the field of video synthesis in a funny way. In the video, Smith surprises viewers by showing eleven video segments in which he actually eats pasta.
With the announcement of OpenAI’s Sora video synthesis model, it is noted that there has been a major breakthrough in the quality of videos produced with artificial intelligence in the last year. Smith’s new video reveals this progress by comparing it to the old AI-generated pasta video. The screen features the AI-generated pasta video from a year ago and segments titled “AI Video Now” in which the real Smith hilariously eats pasta.
Below the video, it appears that some viewers were confused about the new (not AI-generated) video and made comments such as “I’m not sure if the second video was AI-generated either.” This starts a debate about how similar artificial intelligence videos can be to reality.
While we’ve yet to see a model like Sora attempt to reproduce the Will-Smith-pasta-eating video, it’s sure to be a much better result than last year. Considering the rapid advancement of technology, it is predicted that by 2025, video synthesis artificial intelligence models will be able to produce similar ones to the parody video created by Smith.
While last year’s pasta video didn’t represent the most advanced text-to-video synthesis technology at the time it was created in March 2023 (that title likely belonged to Runway’s Gen-2, which was in closed testing at the time), it was still pretty advanced among open-heavy models. This video was produced using the ModelScope AI model, and although more capable video synthesis models were available at the time, the cultural reference to Will Smith grotesquely eating pasta is thought to make for a fun comparison.