AI Video Imagines the LA Olympics in 1,000 Years’ Time

With the Paris Olympics done and dusted, the next Summer Games will be held in Los Angeles in 2028. This gave one filmmaker pause for thought: What would an L.A. Olympics look like 1,000 years in the future?

This being the age of generative AI, creatives can embark on such ambitious projects. So, Josh Kahn turned to AI video generator Runway to create a vision of what the year 3028 might look like on the West Coast.

Immediately, viewers will notice that rising sea levels have pushed the ocean into the city. Presumably, Venice and other coastal areas of L.A. are now underwater. There is then a tour of the various stadiums, all of which are futuristic and ambitiously designed. The swimming arena is in the middle of a lake and the soccer stadium looks an awful lot like the Rome Colosseum.

Kahn tells the MIT Technology Review that while it is stirring to think about what the Olympics of 1,000 years’ time might look like, the video is ultimately a demonstration of AI video’s capabilities.

“We were watching the Olympics and the amount of care that goes into the cultural storytelling of the host city,” Kahn explains to the MIT Technology Review.

“There’s a culture of imagination and storytelling in Los Angeles that has kind of set the tone for the rest of the world. Wouldn’t it be cool if we could showcase what the Olympics would look like if they returned to LA 1,000 years from now?”

AI video works just like AI image generators: The users type in a prompt and the algorithm interprets it to synthesize an image. Waiting for the pictures to appear is a little like waiting for a Polaroid, except with Polaroids the photographer has a pretty good idea of what will appear but with AI it’s far less predictable.

Kahn tells the MIT Technology Review that users of generative AI must be patient and be prepared to go through trial and error. Adding that creating a consistent AI video is very difficult

“These technologies are always better on large-scale things right now as opposed to really nuanced human interaction,” Kahn says.

Kahn predicts that because of this, filmmakers may initially use AI video for wide shots where the magnification of AI’s mistakes isn’t so magnified.