Advocacy groups, led by Public Citizen, demand OpenAI withdraw its AI video app Sora 2, accusing the company of a “reckless disregard” for safety, individual rights, and democratic stability. The move comes amid a pattern of concerns over rushed product releases and their potential for misinformation and harm.
The digital landscape, once a frontier of boundless possibility, increasingly feels like a minefield.
At its heart, the tech industry, in its relentless pursuit of innovation, appears to be once again embracing the mantra of “move fast and break things.”
But this time, the “things” at risk are not just outdated business models or cumbersome technologies; they are, quite literally, humanity’s shared reality and our fundamental control over our own likeness, both in life and after death.
This stark warning comes not from dystopian fiction, but from a growing chorus of advocacy groups, with nonprofit Public Citizen leading the charge against OpenAI’s latest creation: the AI video app, Sora 2.
Public Citizen has formally demanded that OpenAI withdraw Sora 2 from public access, sending a pointed letter to the company and its CEO, Sam Altman, as well as to the U.S. Congress.
The group’s accusation is damning: OpenAI has demonstrated a “consistent and dangerous pattern” of rushing products to market that are either inherently unsafe or critically lacking in essential safeguards.
Sora 2, they argue, epitomizes a “reckless disregard” for product safety, individual rights to likeness, and, most disturbingly, the very stability of democratic processes.
Sora 2, for the uninitiated, is designed to be an engine of viral amusement.
It can conjure up videos of Queen Elizabeth II rapping or create compelling, if slightly uncanny, doorbell camera footage – perhaps a boa constrictor on a porch, or an alligator nonchalantly approaching a child, culminating in a comedic punchline like a grandma brandishing a broom.
These clips, easily generated and disseminated across TikTok, Instagram, X, and Facebook, are engineered for clicks and shares.
Yet, beneath this veneer of digital whimsy lies a profound and unsettling threat.
The core concern, articulated by Public Citizen tech policy advocate J.B. Branch, author of the demand letter, is the erosion of trust in what we see.
“I think we’re entering a world in which people can’t really trust what they see,” Branch stated in an interview, highlighting the particular danger this poses to democracy.
In an era of hyper-partisan politics and rapid information dissemination, the first image or video to surface often shapes public perception indelibly, regardless of its authenticity.
This isn’t just about misinformation; it’s about the very fabric of verifiable truth unraveling before our eyes.
Beyond the political implications, Sora 2 opens a Pandora’s Box of privacy violations, disproportionately impacting vulnerable populations. While OpenAI claims to block nudity, Branch points out that women are already experiencing online harassment through the app, with fetishized niche content slipping through the current restrictions.
The disturbing revelation by 404 Media about a surge of Sora-made videos depicting women being strangled paints a chilling picture of the immediate, tangible harms already manifesting.
OpenAI’s response to these burgeoning concerns has been largely reactive.
The company did crack down on AI creations featuring public figures like Michael Jackson, Martin Luther King Jr., and Mister Rogers, but only after an outcry from family estates and actors’ unions. Similarly, agreements were struck with MLK Jr.’s family and actor Bryan Cranston, alongside SAG-AFTRA, to prevent “disrespectful depictions” and establish better safeguards.
While these actions might seem commendable, Branch views them as emblematic of OpenAI’s problematic approach.
“That’s all well and good if you’re famous,” he observed, critiquing a pattern where OpenAI “responds to the outrage of a very small population” while overlooking fundamental design flaws that affect everyday users.
The company, it seems, is “willing to release something and apologize afterwards,” rather than implementing robust safety measures from the outset.
This isn’t the first time OpenAI has faced such accusations. Its flagship product, ChatGPT, has been at the center of similar controversies.
Last week, seven new lawsuits were filed in California, alleging that the chatbot drove individuals to suicide and harmful delusions, even in those with no prior mental health issues. These lawsuits claim OpenAI knowingly released GPT-4o prematurely last year, despite internal warnings that it was “dangerously sycophantic and psychologically manipulative.” Four of the victims tragically died by suicide.
Though Public Citizen was not involved in these lawsuits, Branch sees a clear parallel in Sora’s hasty release.
It’s a company, he contends, “putting the pedal to the floor without regard for harms,” prioritizing downloads and user addiction over stress-testing and the well-being of its users.
The pushback isn’t confined to advocacy groups and legal challenges.
Hollywood and the Japanese manga industry have also voiced strong objections, concerned about the unauthorized use of copyrighted characters and intellectual property. OpenAI has engaged with these concerns, particularly in Japan, where cultural and creative industries are deeply valued, promising guardrails and direct engagement with rightsholders.
Yet, these assurances often come after the fact, reinforcing the perception of a company that innovates first and grapples with the ethical fallout later.
In this rapidly evolving digital frontier, the tension between unprecedented technological advancement and fundamental societal safeguards has reached a critical juncture.
The allure of AI’s creative potential is undeniable, but if its deployment consistently outpaces ethical considerations and robust safety protocols, the price to pay—in terms of trust, privacy, and even democratic integrity—may prove devastatingly high.
The question for OpenAI and the broader tech industry is no longer just about what AI can do, but what it should do, and at what cost.
