OpenAI recently released Sora, a new deepfake generator that lets users generate custom personalized videos of themselves. This age-defying app is rolling out a really fun feature of “cameo.” It empowers consumers by allowing them to determine who is permitted to generate videos that leverage their likeness. With a focus on user permissions, Sora does not permit the generation of videos of real individuals without their explicit consent, although it applies more lenient standards for deceased historical figures.
Sora’s capabilities extend beyond user-generated content. The app produces alarmingly lifelike deepfakes of world leaders. This means their lives were shaped not just by politicians like Donald Trump, but historical icons like Abraham Lincoln. OpenAI has fine-tuned the video generator to ensure adherence to the laws of physics. WaPo generated this output, which is convincing and stunningly captivating. The app’s content feed has been flooded with more quirky short-form videos starring animated cartoon characters, including Pokémon’s Pikachu and SpongeBob SquarePants.
The “cameo” feature in particular is worth a mention, as it gives users control over their cameo privacy settings. They can choose from various options: “only me,” “people I approve,” “mutuals,” or “everyone.” This uncapped, user-resistant feature assumes full control over users’ digital likeness and interactions lack any substantial, meaningful protection. It further increases the application’s attractiveness in an age when consent and privacy is a priority.
Sora’s advanced and unusual capabilities have excited animal enthusiasts as well as worried watchdogs. Though the app does provide a new medium for artistic creativity, it contributes to the ongoing debate about the ethics of deep fake technology. OpenAI has introduced parental controls to protect users under age 18, creating a safer space for all of the platform’s users.
While Sora’s rules can be very tight regarding living characters, they have more freedom when it comes to dead figures. Consequently, creative possibilities are abounding. One of the most fun examples stars OpenAI CEO Sam Altman, surrounded by a colorful landscape of Pokémon. This should illustrate (figuratively) the app’s new and powerful ability to create whimsical menagerie scenes. Videos depicting Altman engaging in various activities, such as ladling GPUs into bowls at soup kitchens, have emerged, showcasing the versatility of Sora’s deepfake technology.
“I grew up in Philadelphia, so the Phillies are basically the soundtrack of my summers.” – my AI deepfake
While users discover all of Sora’s features, they’ll find even more personalized content informed by their unique preferences. By using IP geolocation and ChatGPT history, the app is able to tailor responses to each user’s specific context, delivering a high-value personalized experience. This level of integration of generative AI technology dramatically transforms and maximizes the app’s functionality and ability to engage users.
As with any new tech, the fear of use and potential impacts are still present. Sora in its own materials has indicated that AI output might still breach its content creator guardrails related to third-party likenesses. As the app goes through further iterations and updates, ethical implications will be an ongoing consideration for OpenAI.
OpenAI’s new approach to deepfakes with Sora is a historic moment for all digital content creation. It’s an idea that enables the company’s customers to develop their own digital avatars. Simultaneously, it maintains ethical standards, redefining the best practices for the industry.