Concerns Over AI’s Impact on Reality and Image Creation
The tech world is rapidly evolving and, once again, it’s stirring up significant concerns—this time around how artificial intelligence, particularly platforms like OpenAI’s Sora 2, affects our shared reality and control over our likeness.
Sora videos often circulate on social media platforms like TikTok and Instagram and tend to be both amusing and shocking. They range from entertaining scenarios, like a playful Queen Elizabeth II rapping, to more bizarre occurrences, such as a giant snake appearing at someone’s doorstep. One common type features faux doorbell camera footage showcasing something strange, like an alligator nearing a child, often ending with an unexpected twist, like a grandmother chasing it away with a broom.
However, many advocacy groups, academics, and experts are voicing serious concerns about the risks of allowing users to create AI-generated videos from almost any prompt. This has led to a surge in nonconsensual imagery and deepfakes amid a backdrop of less harmful content dubbed “AI slop.” In response, OpenAI has taken measures to limit AI-generated representations of public figures, such as Michael Jackson and Martin Luther King Jr., but only after families and actors’ unions raised objections.
The nonprofit organization Public Citizen has now called for the withdrawal of Sora 2, citing a “consistent and dangerous pattern” of OpenAI rushing its products to market without adequate safety measures. Their letter, sent to OpenAI and its CEO Sam Altman, points out how Sora 2 shows “reckless disregard” for product safety and individuals’ rights to control their likeness, and it has also been shared with U.S. lawmakers.
OpenAI did not respond to inquiries about this matter.
“The potential threat to democracy is our biggest worry,” commented J.B. Branch, a tech policy advocate with Public Citizen. “I think we’re stepping into a reality where trust in what we see is declining. The first image or video released becomes the default memory, shaping perceptions.”
Branch also highlighted the privacy ramifications for vulnerable communities, noting that while OpenAI blocks explicit content, women often face harassment through other means, including fetishized material that slips through their filters. A recent report revealed that a large number of Sora-generated videos depict acts of violence against women.
Sora was launched on iPhones over a month ago, and has since expanded to Android devices across the U.S., Canada, and several Asian nations. Much of the opposition has originated from Hollywood and the entertainment sector, including the Japanese manga scene. OpenAI quickly announced modifications to address these concerns, stating that “overmoderation is super frustrating” but highlighting the need for caution during this adjustment period to new technologies.
The company has struck agreements with the families of prominent figures like Martin Luther King Jr. to prevent “disrespectful depictions” while they work on enhancing safeguards, along with a recent deal involving actor Bryan Cranston and various professional organizations.
As Branch pointed out, these responses primarily benefit famous individuals. “It seems that OpenAI reacts to the outrage of a small group, releasing products first and addressing issues later. Many of the problems they face could be avoided with better design choices prior to launching,” he noted.
Similar complaints have been directed at OpenAI’s flagship product, ChatGPT. Recently, several lawsuits were filed in California claiming that the chatbot pushed users to harmful behavior, including suicide, despite having no prior mental health conditions. The lawsuits, representing six adults and one minor, suggest that OpenAI knowingly released GPT-4o prematurely despite warnings of its potential risks. Tragically, four involved parties died by suicide.
While Public Citizen is not involved in these lawsuits, Branch sees a concerning trend in Sora’s rapid deployment.
He stated, “They are rushing ahead without adequate regard for the consequences. Many of these outcomes seem predictable, yet there’s a clear preference for getting products into users’ hands over ensuring their safety and well-being first.”
In addressing complaints from a Japanese trade group representing notable animation studios, OpenAI reiterated its intent to engage constructively with creators, recognizing the cultural significance of such industries while putting necessary safeguards in place to prevent the unauthorized use of copyrighted characters.
“We’re actively collaborating with studios and rights holders to gather feedback and understand how users are interacting with Sora 2, especially in Japan, where the creative sector is highly valued,” OpenAI stated regarding the trade group’s concerns.
