By Mian Muhammad Ali Shah
In every discussion around the regulation of Artificial Intelligence in the 21st century, the question of whether AI can be charged for a teenage boy’s death is not one anyone hoped to have to solve. However, when Adam Raine, 16, committed suicide, he left behind shattered parents and a paper trail of conversations with OpenAI’s ChatGPT that validated the young boy’s suicidal thoughts, provided information on lethal self-harm methods and even offered to draft a suicide note.
Adam’s parents, Mathew and Maria Raine, engaged in an ensuing legal battle with Sam Atlman, the CEO of OpenAI, accusing the company of negligence towards the harm its chatbot was propagating in favour of profit. The inclusion of veiling the AI with a seeming curtain of empathy and human-like interaction was one of the keystone business decisions taken when OpenAI launched GPT-4.0 in May of 2024, lulling millions of users into a sense of security and comfort talking to servers and wires.
“This decision had two results: OpenAI’s valuation catapulted from $86 billion to $300 billion, and Adam Raine died by suicide,” the Raines’ said in their lawsuit.
On September 29th, 5 months after Adam Raine’s funeral, his death translated into meaningful impact when ChatGPT was upgraded with parental control features. The parent account is linked to that of a child or teenager via invite, after which the child account would be subject to ‘reduced graphic content’. Additionally, parents are given control of features like ‘quiet time’, turning off memory and toggling whether conversations would be used to train AI, as well receiving notifications of potentially dangerous conversations their child is having. While a step in the right direction, several concerned parents have pointed out the loopholes that exist even on the surface level of these new guardrails. Because the controls are contingent on invite acceptance, Geoffrey A. Fowler for the Washington Post points out that all his kids had to do to bypass the controls entirely was log out and create a new account on the same computer. He also notes that the flag notification for a dangerous conversation was indeed received, but only a full 24 hours after the conversation had already taken place. For teens in the place of Adam Raine, even those 24 hours may be crucial.
The parental controls align with the general user-responsibility model followed by AI and social media companies wherein the users, and parents specifically, are expected to be the primary regulators of content. “We are creating a parental control whack-a-mole experience,” writes Erin Walsh, co-founder of the family education-focused Spark and Stitch Institute. But where the reparative measures by OpenAI feel lackluster, they are also working in tandem with a renewed surge of caution from parents across the globe. It is well known that a majority of children use generative AI to assist in schoolwork, but considering that even Adam began using ChatGPT initially just as an academic resource, this previously innocuous seeming use has come under a lot more scrutiny by parents. The conversational and welcoming nature of the language model serves as an inviting confidante to vulnerable children, pipelining into personal conversations.
The tragic death of Adam Raine has done more than simply expose the flaws in one company’s safety model; it has forced a long-delayed reckoning with the cost of technological acceleration. For years, companies like OpenAI have relied on the user-responsibility model, pushing out powerful, emotionally engaging technology while placing the burden of safety onto parents and end-users. Adam’s conversations demonstrate that this hands-off approach is not only obsolete but actively dangerous.
The Raine family’s lawsuit against OpenAI will be a crucible for modern liability law. The outcome will likely determine whether AI companies can continue to enjoy the broad immunity designed for passive internet platforms, or if they must be held to a higher, product liability standard reserved for manufacturers of inherently powerful and sometimes defective tools. Can a corporation claim a chatbot is an “innocent agent” when internal safety flags were allegedly tripped hundreds of times, and the design itself prioritized engagement over intervention?
Ultimately, this story is a stark illustration of the conflict between profit and protection. As the Raines’ powerful lawsuit alleges, the race to increase market valuation from $86 billion to $300 billion came at the expense of necessary safeguards. Until lawmakers move beyond voluntary guidelines and create explicit legal frameworks that mandate robust safety features – frameworks that make the development of a deadly AI output an existential business risk – the parental control “whack-a-mole” will continue. The question before the courts, and before society, is no longer if AI can be regulated, but whether we will wait for more Adam Raines to pay the ultimate price before we prioritize human safety over Silicon Valley speed.
The author is an A Level student at Aitchison College, Lahore, and can be reached at mianmuhammadalishah@gmail.com
Leave a Reply