AI has now blurred the threshold between reality and obvious fantasy, and that is an issue. With the introduction of Sora 2, backed by OpenAI, the action of typing out fictional prompts and larger-than-life scenarios was enhanced.
This app, along with a plethora of others, has generated breathtaking videos. Some videos feature Polaroid-style prompts of an older and younger version of yourself, a beautiful trend that has made its way to multiple platforms. Some include point-of-view videos of you, the viewer, battling fictional characters. Imagine you versus Batman, Superman, or whoever you want.
With a few clicks and typing sequences, anything you would like to become reality will do just that, and with the prevalence of AI-based apps increasing, AI has been becoming stronger and stronger. TikTok has been surging in these types of videos, with viewers constantly struggling to determine whether the content is real or fake.
Viewers are struggling to distinguish truth from fiction when it comes to AI. That is the problem.
With these apps making their way to more platforms and the content becoming increasingly realistic, they serve as a dangerous platform for people to misinform, to frame and to accuse.
Sora 2 is a level-up from the preceding “Sora” app; matching dialogue, as well as enhanced facial expressions and high-definition movements, are the primary attractions, further creating constant virality within social media audiences. It also features ambient sounds, as well as more precise cameos.
But this also makes for a deepfake frenzy, where citizens and celebrities can have their voices used, as well as their faces, body frames and more, without their consent or permission. Now, a common question about this is, have there been any changes or solutions to this?
OpenAI has backtracked on its initial “opt-out policy,” allowing it to use this information at its discretion, unless those whose info was used demand removal or take legal action. Aaron Moss, lawyer, editor, publisher and blogger of Copyright Lately, has already seen the restrictions and limitations on how much freedom can be used with trademarked characters.
“Testing the app this weekend, I saw just how quickly OpenAI can tighten controls when motivated. Characters from “Family Guy,” “South Park,” “King of the Hill,” and numerous other copyrighted properties that I easily generated on Tuesday now trigger content violations,” Moss said.
Now that this policy has been reversed, rightsholders will be allowed to hand out permission for OpenAI to use their copyrighted content.
Said best by OpenAI CEO Sam Altman in a self-written blog regarding the very first Sora update, he states, “First, we will give rightsholders more granular control over generation of characters, similar to the opt-in model for likeness but with additional controls.”
Elaborating further with Altman’s explanation, rightsholders will have a say so with how their copyrighted content will be displayed. Even with that being placed in, that does not help with the unnecessary violence and graphic nature that is able to be generated with Sora 2.
Some extreme examples could be a video of the 37th president, Richard Nixon, addressing the moon landing as fake to a televised audience, or multiple videos of fictional mass shootings and bomb scares.
This serves as a breach of some guidelines set by OpenAI. According to NPR, their team was able to access loopholes within the app, generating violent prompts, despite OpenAI’s strict policies towards it. “NPR was able to get it to produce short videos on topics related to chemical, biological, radiological and nuclear weapons in direct contradiction of OpenAI’s global usage policies.”
Even with their policies standing firm on preventing these types of generations, it still happens and it is still able to be produced and published.
While Sora 2 may serve as a prompt for larger-than-life videos that create superb fictional scenarios and content, it is more infamous for becoming a setting for accusatory visualization, graphic violence and much more. My primary issue with this app is its contradictory structure. Although many of OpenAI’s policies prohibit violent prompts and copyrighted character usage, they are still being generated by users.
In a world where many citizens can barely tell the difference between real-life content and AI, it’s important to decipher the differences between these videos and what the creators of these apps can contribute to safe measurements because, as far as social media and OpenAI are concerned, AI is going nowhere anytime soon.
