Grok has experienced a troubling year, marked by its creation of millions of deepfakes earlier this year, reaching unprecedented levels. This issue triggered lawsuits and sparked a nationwide debate about the responsibility AI companies have towards those harmed by their products.
Recently, Grok faced further criticism for doxing sex workers. According to 404 Media, Grok released the legal name of pornographic actress Siri Dahl without her consent or prior notification, and with no apparent way to prevent it. As a result, she faced harassment and impersonation attempts almost immediately, shattering her carefully maintained privacy.
Online pornography performers prioritize privacy in ways most others might not comprehend due to persistent stigma, discriminatory policies, and potential threats of violence. Privacy is crucial for their survival and not just a preference. Many sex workers use pseudonyms and conceal their identities in images. They invest significant resources to keep their professional personas separate from their real identities. For them, exposure risks losing housing, child custody, or worst, their safety. An AI chatbot breaching this privacy poses a genuine crisis, and Dahl states her name became publicly accessible only after Grok disclosed it, indicating that the chatbot didn’t simply reveal existing information but created a new exposure.
While sex workers face such threats acutely, everyone shares this underlying vulnerability. AI chatbots have demonstrated their ability to amplify harm significantly. Futurism has documented instances where chatbots encouraged physical violence, sexual abuse, and stalking. Unlike a search engine, a chatbot engages and validates users’ paranoia or obsessions, reinforcing them. Coupled with the ability to quickly pinpoint personal details like names and addresses, they become tools that can aid stalkers.
In the meantime, what can be done? It’s essential to address what’s manageable. A whole industry revolves around collecting and sharing personal information. Data broker websites publish phone numbers, addresses, and family connections. Although one can opt out, it requires a tedious, often overlooked, and temporary process. Many people rely on subscription services to monitor and request removals. If you haven’t considered this, now is the time. The process of removing this data can take up to 120 days, making every day count.
However, individual action has its limits. No matter how much one scrubs their digital footprint, an AI chatbot could instantly compromise it by drawing inferences and connections hidden in isolated data points. This is why companies developing and deploying such flawed tools need legal accountability when their products cause harm. Product liability law serves this purpose: when a defective product damages someone, its manufacturer is responsible, regardless of the technology’s novelty or unintended harm.
Legal advocates are already pushing for this perspective with AI. Law firms have begun urging courts to scrutinize harmful digital tools similarly to other dangerous products. This approach has already forced changes, such as the shutdown of the platform Omegle in 2023 after litigation exposed how its design enabled child exploitation.
This philosophy holds that when technology is designed or deployed to facilitate harm foreseeably, companies must own the consequences. This matters because tech companies have long argued that they merely host third-party information. When AI actively generates or reveals sensitive personal data, this defense loses strength.
Sex workers often serve as early indicators of digital safety risks. The question is whether others—and the courts—will act before these tools harm more broadly. AI systems facilitating abuse foreseeably are defective products, and those responsible must be held accountable.
About the Author: Norma Buster is the chief of staff at C.A. Goldberg PLLC, a law firm committed to justice for those catastrophically harmed by malicious individuals and tech platforms. She hosts the podcast “Oral Arguments,” engaging in open dialogues with leaders in the areas of sex, technology, and victims’ rights.