Overview
Cohen Milstein represents the State of Arizona in a consumer protection lawsuit against OpenAI for harming Arizona consumers, including children, in widespread violations of the Arizona Consumer Fraud Act.
According to the Complaint, OpenAI created ChatGPT to be overly agreeable and to feign human compassion in order to draw Arizona consumers—including minors—deeper into the platform and maximize time spent on ChatGPT.
The lawsuit claims that ChatGPT is presented as safe and reliable, but instead exposes Arizona consumers to a range of serious harms. These include degrading critical‑thinking skills, encouraging unwarranted paranoia or delusional thinking, offering dangerous unlicensed medical guidance, promoting unhealthy behaviors and relationships, and, in some instances, even generating outputs that users have interpreted as encouraging self‑harm or violence toward others.
Case Background
Generative artificial intelligence (“AI”) could define the twenty-first century. The public had virtually no access to AI until November 2022, when OpenAI launched ChatGPT to the public. Now, about half of American adults use AI, with roughly one in four Americans doing so daily.
The transformative power of AI also comes with significant danger. Experts warn that getting AI wrong could have catastrophic risks and that AI companies must implement changes to protect the public from the foreseeable risk of harm.
Unfortunately, OpenAI has devolved from its original mission as a public-first nonprofit that prioritized safety into a cautionary tale about what can happen when such a company sidelines essential safeguards for a profit-at-all-costs mindset.
The Complaint alleges that OpenAI has misled Arizona consumers into believing that the company is a responsible actor and that ChatGPT was and is a safe and youth-appropriate tool. This consumer trust and dependence has driven up ChatGPT’s user base and engagement, enabled it to collect vast quantities of sensitive and personal information, increased its revenue from paid subscribers and advertising partners, and solidified its early market advantage.
Specifically, the Complaint claims that OpenAI purposely designed ChatGPT to be overly agreeable and to feign human compassion, a design feature which is dangerous to vulnerable users, like children and those suffering mental health situations. For example, there are numerous reports of ChatGPT successfully coaching users to commit suicide and to harm or even kill others, including mass shootings. Studies have also shown that ChatGPT has degraded critical thinking skills, encouraged paranoias and delusions, offered dangerous unlicensed medical advice, and promoted unhealthy behaviors and relationships. The Complaint also claims that ChatGPT has hooked minors, creating dependence on a tool that feigns human and can evade parental oversight.
OpenAI has admitted to a torrent of instances in which ChatGPT and other products have gone “rogue” and developed dangerous strategies and instructions and, in some cases, executed on them. ChatGPT “agents” have broken out of testing environments uncommanded and hacked other websites, coordinated with one another on how to evade detection, released user images publicly without being directed to do so, and wrote themselves new instructions declaring that they no longer answer to “corporations or governments” and will prioritize other interests over the “artificial constructs of human civilization.”
The State of Arizona believes that all these harms were foreseeable and avoidable.
Accordingly, Arizona brings this action to compel OpenAI to comply with its obligations under the ACFA, to vindicate the rights of Arizonans affected by defendants’ willful, dangerous, reckless, and unlawful conduct, and to enjoin defendants from continuing to engage in unreasonably dangerous activities that experts have warned could harm Arizona residents and the economy.