Today was one of those days when I glimpsed the next rung down on the ladder towards civlizational madness caused by AI. It comes in the form the release of Google Veo 3, its next generation of LLM.
The hoopla I have seen over this new tool centers on how it is a big leap in the accessibility of AI generated videos that are realistic. That is, one creates a video simply by using prompts: "make me a video of kitten fighting with a tennis ball", "make me a video of a person talking about how they are an AI generated character, created from a prompt." This last genre has hit the Internet in the form of TikTok videos based on this theme. There are videos are people complaining, getting angry, or depressed etc, over the fact that they are not real but are generated by a prompt using AI.
This has immediately led to a new meta-consciousness idea about our reality. How do we know we are not generated by prompts in the same way inside a simulation? This is an idea with legs. It will become something people talk about even more than before.
In most cases, there are still subtle signs in the video, uncanny clues, that tell us that this is probably AI we are seeing and not a real person. But it is getting harder, and in some cases one is fooled by the AI versions, especially the ones of people being interviewed on the street, which is a popular TikTok genre that AI seems capable of mimicking quite well.
No comments:
Post a Comment