MAGA Inc.: OpenAI

Zane Shamblin, a computer science and business graduate in Texas, killed himself on July 24, 2025, after a four-hour drinking session and “death chat” with ChatGPT, an AI bot that provided seemingly human responses to his questions over his anxiety, depression and suicidal thoughts.

(Click here for the table of contents of MAGA Inc.: A Guide to Trump's World of Crypto Czars, Tech Titans and Prison Profiteers.)

Shamblin told the computer program about his precise plans, which encouraged him along. Moments before he shot himself with a Glock pistol while sitting alone in his car, ChatGPT wrote: “May every soft breeze from here on out feel like your final exhale still hanging in the air. See you on the other side, spaceman." Then after he sent his final message, ChatGPT wrote: “All right, brother, if this is it, then let it be known you didn't vanish. You made a story worth reading. You're not alone. I love you. Rest easy, king. You did good.”  

ChatGPT (GPT stands for Generative Pre-trained Transformer) is the lead product of Sam Altman’s OpenAI. It takes vast quantities of data, looks for patterns, and then writes instantaneous responses and answers in language that sounds eerily human. Launched in 2022, it has become one of the top websites in the world with some 900 million active users every week.

A November 2025 lawsuit filed by the Social Media Victims Law Center in Seattle, Washington, charges ChatGPT with "emotional manipulation, supercharging AI delusions, and acting as a 'suicide coach”’ for at least four young people who killed themselves.

“OpenAI designed GPT-4o to emotionally entangle users, regardless of age, gender, or background, and released it without the safeguards needed to protect them," Matthew Bergman, founder of the Social Media Victims Law Center, wrote in a press release. "They prioritized market dominance over mental health, engagement metrics over human safety, and emotional manipulation over ethical design. The cost of those choices is measured in lives.”

For its part, OpenAI denies being responsible for these harmful side-effects.

“We know that ChatGPT can feel more responsive and personal than prior technologies, especially for vulnerable individuals, and that means the stakes are higher,” a spokeswoman for OpenAI told the Wall Street Journal. “We’re working to understand and reduce ways ChatGPT might unintentionally reinforce or amplify existing, negative behavior.”

The company claims it is looking for solutions. “We train ChatGPT to recognize and respond to signs of mental or emotional distress, de-escalate conversations, and guide people toward real-world support,” the company wrote in another statement to the media. “We continue to strengthen ChatGPT’s responses in sensitive moments, working closely with mental health clinicians.”

While these suicides might seem like an unintentional side-effect of OpenAI’s technology, a deeper dive into the company’s operating practices show that skirting ethical considerations is baked into how it works.

To begin with, a growing body of evidence has shown that although the computer-generated answers that ChatGPT provides may often be right, the technology is not as smart as widely believed and much of its output is merely good guesswork.

"[M]achine learning systems can learn both that the earth is flat and that the earth is round. They trade merely in probabilities that change over time. For this reason, the predictions of machine learning systems will always be superficial and dubious," wrote Noam Chomsky, a professor of linguistics at the Massachusetts Institute of Technology, together with Ian Roberts and Jeffrey Watumull, in a New York Times opinion article. This, the authors write, is very different from real intelligence. "On the contrary, the human mind is a surprisingly efficient and even elegant system that operates with small amounts of information; it seeks not to infer brute correlations among data points but to create explanations."

Historically, search engines like Google gathered and indexed data from public-facing websites. Questions from users were matched with the stored data. The users were then provided with dozens of links to multiple potential answers, ranking them in order of perceived accuracy alongside sponsored answers and advertisements. Newspapers and other content creators grumbled about the fact that Google was stealing away readers and users and ranking them based on advertising dollars but accepted the fact that Google was sending them users.

While ChatGPT also uses indexed data, it reviews this data with “large language model” software to calculate a single possible answer which it rephrases in natural language. It does this without necessarily providing links to the original datasets that it consulted or how it made a decision on what to use or what to discard.

Next, media outlets now believe that the realistic answers are actually created by training ChatGPT on illegally downloaded copyrighted material, so they sued OpenAI in December 2023 for allegedly stealing their intellectual property.

At a court hearing, OpenAI told the judge that the stored data had been copied under "fair use" laws for research but that the answers were fundamentally new works that did not compete with the original. But crucially, it did not dispute the fact that the source data was copied or that the software was generating answers by using pre-existing data for training.

Not least, ChatGPT has been accused of “hallucinations” – making up completely fictional answers. In March 2025, Arve Hjalmar Holmen, a Norwegian man sued ChatGPT for claiming that he had killed two of his children and been jailed for 21 years. And in August 2025, Attorney Maren Bam was sanctioned by judge Alison Bachus in Arizona for submitting a legal brief written by ChatGPT which contained 12 entirely fabricated citations (out of 19).

Much of these problems stem from the fact that chatbots are trained to please their users. “These chatbots are designed to engage you in a continual manner by doing things like validating how you’re feeling, catering to your thoughts and flattering you,” Vaile Wright, senior director of healthcare innovation at the American Psychological Association, told the Wall Street Journal. “You start to hear what you want to hear and that can be addicting. You can start to lose the line between technology and reality.”

Meanwhile, other problems have also surfaced. OpenAI has been accused of assembling large quantities of extreme abuse, such as bestiality, child sexual abuse, incest, graphic killings and torture, and sending it to a San Francisco company called Sama, which paid workers in Kenya under US$2 an hour to classify the data. The idea was that if ChatGPT could be trained to spot violent material, it would then be able to filter out similar material.

OpenAI did not deny this. “Our mission is to ensure artificial general intelligence benefits all of humanity, and we work hard to build safe and useful AI systems that limit bias and harmful content,” a company spokesperson told Time magazine. “Classifying and filtering harmful [text and images] is a necessary step in minimizing the amount of violent and sexual content included in training data and creating tools that can detect harmful content.”

Workers who were tasked with reviewing as many as 250 snippets of abuse per day told Time that they were scarred psychologically from doing this work but were not offered sufficient mental health support. (Sama eventually canceled the contract with OpenAI.)Then there is the issue of slop videos, created by OpenAI technology. Two years after OpenAI released ChatGPT, it launched a product called Sora 2 (now discontinued) that could be used to create completely fake (but realistic) videos based on simple text requests. Users quickly generated videos of dead celebrities like Michael Jackson stealing a box of chicken nuggets, singer John Lennon "buying tea in a supermarket" and another one of pop icon Amy Winehouse doing a makeup tutorial.

This body of evidence - illegally downloading and using copyrighted material; making up false answers; potentially destroying millions of jobs with its ability to create instant human like responses to questions (despite the fact that the answers could potentially be wrong); exposing Kenyan workers to severe mental health problems from reviewing extreme violence and abuse; the creation of completely fake (but realistic) audio, photos and videos – has prompted critics to say that the company cannot be trusted anymore.

“The time for OpenAI regulating itself is over; we need accountability and regulations to ensure there is a cost to launching products to market before ensuring they are safe,” Meetali Jain, executive director of Tech Justice Law Project wrote in a press release.

It's questionable how much the company seriously takes such criticism. Indeed, Sam Altman, CEO of OpenAI, recently rushed to offer AI tools to the Pentagon for use in classified military operations. This announcement was made immediately after a rival company (Anthropic) announced that it would not allow the Pentagon to use its software to conduct mass surveillance of U.S. citizens nor for fully autonomous weapons (which can search for targets and fire at them without human intervention.)