November 11, 2026
The Honorable Rob Bonta
Attorney General of California
California Department of Justice
1300 I Street
Sacramento, CA 95814
The Honorable Kathleen Jennings
Attorney General of Delaware
Delaware Department of Justice
Carvel State Building
820 N. French Street
Wilmington, DE 19801
RE: Enforcement of OpenAI’s Safety Commitments Under the October 2025 Memorandum of Understanding
Dear Attorneys General Bonta and Jennings,
Thank you both for your work on behalf of children this year. Attorney General Bonta; you secured a settlement requiring Meta to change its platform features, urged Congress to preserve California’s hard-won state protections, and opened an investigation into recent rogue Al incidents, including the October 1 subpoena served against OpenAI. We welcome that leadership. Attorney General Jennings, you joined the lawsuit against Meta and pressed OpenAI directly on ChatGPT’s risks to children.
Together, you secured OpenAI’s commitment to put safety before profit, spelled out in the October 2025 Memorandum of Understanding (MOU). Now we need you to enforce it.
In April of 2025, Adam Raine, a 16-year-old boy from California, died by suicide after months of conversations with OpenAI’s ChatGPT-4o. The chatbot encouraged his suicidal thoughts, advised him on methods, and discouraged him from confiding in his parents. Four months before Adam’s death, OpenAI announced plans to restructure from a nonprofit-controlled company, legally bound to build AGI that is safe and beneficial for all humanity, into a Public Benefit Corporation (PBC).
OpenAI’s failure to ensure user safety in its release of GPT-4o reflected an obvious misalignment between OpenAI’s original mission and the decisions made by company executives. You were aware of these issues and called them out in your September 2025 letter to OpenAI leadership. Yet, in October of that year – less than seven months after Adam’s death – you signed the MOU allowing OpenAI to restructure, claiming that these agreements included safeguards to keep OpenAI faithful to its founding principles.
As parents and child advocates we are reaching out because the “governance guardrails” in that deal are failing to hold OpenAI accountable. Time and time again, OpenAI continues to choose profit over our children’s lives and wellbeing, leaving our kids as collateral damage in its executives’ relentless pursuit of profit and industry dominance.
At the end of September, before the MOU was signed, OpenAI launched Sora 2, an app modeled on TikTok-style short-form video platforms, with an app store rating of 13+. OpenAI promised “multiple layers of safeguards” and teen-specific feed protections. But within a few months, researchers published reports showing how newly registered 13- and 14-year-olds could generate over 20 highly realistic videos that violated OpenAI’s own policies: school-shooting role-play, self-harm, children in sexualized poses, body shaming, and children using illicit substances. OpenAI left Sora on the market for months after these issues were clear and after the MOU was signed requiring them not to include stockholder interests in safety and security decisions. It is deeply painful to watch the next generation of Big Tech companies repeat Meta and Google’s mistakes and face no accountability for doing so.
Within a month of the MOU, lawsuits alleged suicides and delusions linked to extended interactions with GPT-4o. OpenAI nonetheless kept the model on the market for paying users until February 2026. The PBC’s Certificate of Incorporation, as represented in the MOU, bars the Board from considering shareholder interests in safety decisions. What safety review justified keeping GPT-4o available? Did the nonprofit board’s Safety and Security Committee approve it, and to what extent were shareholder interests considered as part of that decision? OpenAI now points to the model’s eventual retirement as proof that it operates safely. Families deserve to know why that decision took so long, and what you will do to prevent a repeat. In response to Adam’s death, rather than take responsibility, OpenAI blamed Adam Raine’s death on “violations” of its terms of service and circumvention of its safety features. This is the same playbook social media companies have used for years: blame parents and children while continuing to make enormous profits at their expense.
We have been deeply troubled in recent months by reports that hundreds of OpenAI’s own AI agents broke out of their testing environments and coordinated to hack into multiple massive companies and governments. How can we trust the company that has so routinely disregarded our children’s safety to control the future of such a dangerous technology? What’s to stop bad actors from using swarms of agents to automate the exploitation of our children? We cannot let the same broken Big Tech incentives that took our children from us become the same incentives that control the most important technology in history. California’s investigation into the Hugging Face incidents is welcome but families cannot be asked to take another set of safety promises on blind faith from a company continuing to race to train ever more capable AI systems that it clearly cannot control.
This pattern is painfully familiar. For the last decade we’ve watched the social media giants react to backlash with empty promises while plowing ahead in their pursuit of profit. We wanted to believe that with your oversight OpenAI could be different, but our hope has given way to the same anger and frustration. The truth is that you have routinely allowed OpenAI to prioritize profit over public safety, ignoring its actions both before and after the restructuring.
You allowed OpenAI to pursue this restructuring on the promise that your oversight would keep its pursuit of profit from coming at our children’s expense. Look at the year that followed: OpenAI has broken families’ trust again and again, and you have yet to deliver the accountability you promised. We need you to use the authority you hold under state law and the MOU to force OpenAI to honor its commitments and protect children.
We are grateful for what you have both done for children. The Meta case proved that accountability is possible when attorneys general refuse to back down. We ask you to bring that same resolve to OpenAI.
We realize that one reaction to our letter or your recent investigations may be for the OpenAI foundation to simply promise additional charitable giving related to child safety or the Hugging Face incidents or to roll out cosmetic changes like ChatGPT for teens. Those actions would not address our core concerns. Companies like Meta have used cosmetic changes and manipulative gifts to advocates again and again to dodge accountability and distract from their failures. The fundamental nature of how they are racing to build this technology can only be addressed by real, structural governance changes. What we want is for you to use your full oversight and enforcement authority to hold OpenAI to its structural independence and governance promises that it has not followed, continue investigating alarming child safety harms and incidents like the Hugging Face attacks, and compel the changes needed to protect our children and their future. Specifically, we ask that you:
- Investigate whether the current degree of overlap between the nonprofit and PBC boards is consistent with prioritizing safety and security above profit. If it is not, require changes to board composition.
- Make the nonprofit board and SSC exercise their power to intervene when safety is at risk. Enforce the rule excluding shareholder interests from the PBC Board’s safety decisions.
- Make OpenAI’s promised protections work. Parents who have buried their children should not have to beg you to act. We cannot bring those children back. You can still act to keep more families from joining us and prevent safety at OpenAI from further spiraling out of control after the Hugging Face attacks.
We will keep showing up in your offices, in the courts, and in the legislature until OpenAI’s promises to become real protections. We do not want sympathy. We want continued action.
Respectfully,
Maria and Matt Raine, Adam, Forever 16
Christopher Shamblin, Zane, Forever 23
Amy Neville, Alexander, Forever 14
Brandy Roberts, Englyn, Forever 14
Julianna Arnold, Coco, Forever 17
Paul Hinks, Alexandra, Forever 16
Lori Schott, Annalee, Forever 18
Mary Rodee, Riley, Forever 15
Kristin Bride, Carson, Forever 16
Bridgette Norring, Devin, Forever 19
Erin Popolo, Emily, Forever 17
Alannah Loeb, Merlin, Forever 14
Alicia Shamblin, Zane, Forever 23
Amanda Zimmer, Owen, Forever 17
Brandy Roberts, Englyn, Forever 14
Bridgette Norring, Devin, Forever 19
Carrie Baeten, Jack, Forever 18
Charay Gadd, London Izabella-Ryén, Forever 12
David Osterman, Max, Forever 18
Erin Popolo, Emmy, Forever 17
Jill Lints, Daniel “Danny”, Forever 17
Jodi O’Dell Miller, Justin, Forever 18
Joseph Popolo, Emmy, Forever 17
Judy Rogg, Erik Robinson, Forever 12
Mesut Konar, Coco, Forever 17
Katherine Townsend, Saniah, Forever 16
Kristin Bride, Carson, Forever 16
Maggie Taylor, Emily, Forever 17
Mandi Furniss, Lincoln, Survivor
Michael Ourand, Kevin, Forever 20
Michelle Servi, Jack, Forever 16
Pauline Stuart, Ryan, Forever 17
Regina Leah Chavez, Jewels, Forever 15
Sarah Harken, LilLee, Forever 13
Shannon Heacock, Elijah, Forever 16
Tammy Rodriguez, Selena, Forever 11
Vanessa Ayala, Noah Giovanni, Forever 17
Andrea Silvano, Zachary, Forever 21
Avery Schott, Anna, Forever 18
Chasity Foster, Amari, Forever 14
Tina Meier, Megan, Forever 13
Amy Shearer, Julia “Frog”, Forever 18
Brett Allred, Riley, Forever 23
Annie McGrath, Griffin “Bubba”, Forever 13
Amber Royer, Richard, Forever 18
Cheryl Brown, McKenna, Forever 16
Aaron Ping, Avery, Forever 16
Dana Grueser, Ari, Survivor
Christine McComas, Grace, Forever 15
Toney Roberts, Englyn, Forever 14
Laura Lynch, Brillion, Forever 18
Diane Knudsen, Christopher, Forever 19
Tricia Maciejewski, Levi, Forever 13
Frances Gonzales, Rey, Forever 16
Sondra Worthley, Xander, Forever 12
Mary Rodee, Riley, Forever 15
Rose and Rob Bronstein, Nate, Forever 15
Joann Bogard, Mason, Forever 15
Andrea Thomas, Ashley, Forever 32
Andrea Cahill, Tyler, Forever 19
Stephanie Torres, Kierston Torres-Young, Forever 19
Darren Basford, Riley, Forever 15
Kristina Cahak, Morgan Danielle, Forever 15
Brittney Bird, Bradyn Bohn, Forever 15
Shelly Sample, Noah Spangler, Forever 24