How to take a common-sense approach to video creation tools
A highly recommendable post by Microsoft President Brad Smith on one of his company blogs, “Combating abusive AI-generated content: a comprehensive approach”, addresses the problem with common sense and caution.
The recent advances in algorithmic video generation tools such as Sora, the latest iteration of OpenAI, are by all accounts impressive. While not yet available to users, the web has quickly filled with examples of incredible videos showing Sora’s potential, which the company claims to have published without any modifications, exactly as they came out of the algorithm. While the technology is certainly awe-inspiring, it is also terrifying because of the possibilities it opens up for the creation of all kinds of deepfakes with uses ranging from revenge porn to politics to fake news.
Obviously, from the moment such tools can be developed, it would be pointless to try to prevent this based on fears of possible abuse, a move that would only result in them being available to a few. But if there’s one thing clear at this point, it’s that, as Smith himself says in his article, “we need people to understand that they can’t believe every video they see or every audio they hear.” Of course, this is easier said than done, and requires some compromises of a technological, regulatory, and educational nature.
