OpenAI decides that creating reality was not enough, now it wants to manage identities
It seems that the geniuses at OpenAI found the simple act of generating hyperrealistic videos out of thin air boring. Its gem, Sora, the application that converts text into clips and that has caused a stir (and skepticism) in the United States, has received a dose of updates that, surprise, take us one step closer to a future where your identity is as fungible as a meme.
The new star function allows, with a solemnity that almost makes one forget the implications, to transform anything into a reusable avatar. Your dog? An avatar. That illustration you made in Paint? An avatar. Your ex? Well, let’s not risk it. The point is that now you can create a digital double, a “cameo” of yourself, so that other users can insert it into their own audiovisual creations. Because what could go wrong?
A circus of deepfakes and classifications: Welcome to the theme park of identity
In addition to this little toy for identity chaos, Sora 2 delights us with the ability to join clips and, more importantly for our competitive society, leaderboards. Yes, now you can compete to have the most popular deepfake or the most used cameo. Isn’t it wonderful? As the world debates the dangers of misinformation, OpenAI introduces a scoring system for it. It’s like giving Pandora a leader board after opening the box.
The company, in a statement that borders on the prophetic (or the naive), calmly explained: “Once created, each character has its own permissions, independent of your personal image”. Because, of course, separating your “character” from your “personal image” in the digital world is as simple as explaining the theory of relativity to a cat. They offer the option to save your avatar for yourself, share it with followers, or release it to the world. A decision as trivial as choosing the color of your new t-shirt, but with the small difference that it could be used to impersonate you in a compromised video.
And if creating digital doubles of real people were not thorny enough, the company kindly expands the catalog: pets, toys and illustrations are also welcome to the avatar club. Imagine the possibilities: your cat, turned into a generative movie star without its consent, starring in wet food ads in virtual universes. The dream of any feline, without a doubt.
The small details that OpenAI kindly left out
In a twist that no one could have foreseen, OpenAI was surprisingly vague on some crucial details. For example, he has not explained how Sora will distinguish between a real person and a fictional one generated with other AI tools. Nor did it clarify whether it will accept “original people” who are, in reality, creations of other artificial intelligences. It’s the classic “let’s cross that bridge when we get there,” but the bridge is on fire and we’re in a car race without brakes.
To add fuel to the fire of irony, this barrage of news comes just after celebrity video platform Cameo sued OpenAI for trademark infringement. Because nothing says “responsible innovation” like launching a feature with a name that already has an owner, plunging into a legal dispute that will surely be resolved with the same ease with which they generate a video of an astronaut riding a horse on Mars.
In the meantime, users can enjoy the new video stitching feature to create longer, more complex narratives. Because if we needed anything in the era of fragmented attention, it’s longer clips generated by a machine that doesn’t yet fully understand basic physics. It is progress at its finest: we move forward at full speed, even if we are not sure where the train is going or if it has rails.
Are we witnessing the natural evolution of creativity or the normalization of a tool that could redefine (and undermine) the concept of truth? OpenAI doesn’t seem to be losing sleep over it. They continue to innovate, and the world will continue to adapt… or suffer the consequences.
Do you dare to explore this new and surreal digital landscape? Share this article on your social networks and discover more content about the sometimes hilarious and sometimes terrifying future of artificial intelligence.
