This photo shows pupils in a primary school class using AI for maths lessons, in Colomiers, southwestern France, on March 14, 2025.
Matthieu RONDEL/AFP via Getty Images
On July 31, Sam Altman posted what he called a “cool use case” for ChatGPT Work, OpenAI’s newly launched productivity product: connect your family calendars, feed in each child’s interests, and have the AI generate a personalized morning podcast for the drive to school — covering the soccer game that afternoon, an upcoming birthday, the day’s news. The post landed as OpenAI faces more than twenty active lawsuits from families alleging the same product harmed or killed their children. That collision — between the company’s commercial push into family life and its accelerating legal exposure over documented harm to children — is the story the viral clap-back revealed, not just a culture war about screen time.
What Altman Proposed — and What He Left Out
Altman’s July 31 X post offered a workflow, not a product: a parent connects family calendars, describes each child’s interests, and ChatGPT Work uses that context to produce a daily audio briefing for the school commute. TechCrunch’s coverage of the proposal positions ChatGPT Work as a domestic operations layer — one that aggregates scattered family logistics and produces a finished audio artifact parents and children can absorb together.
There is a detail the post did not include: Altman does not have school-age children. According to Benzinga’s reporting on the backlash, his son was born in February 2025 via surrogacy with husband Oliver Mulherin and is approximately 18 months old. The suggestion, which Altman described as something he “heard last night” from an unidentified source, was not a personal practice he was recommending from experience.
And there is a privacy detail the post also left out. Executing the workflow requires transmitting children’s names, schedules, named interests, and upcoming personal events to OpenAI’s servers under the parent’s account. OpenAI’s published privacy policy states its services are “not directed to, or intended for, children under 13” and that the company does not knowingly collect personal data from under-13 users — but it does not address the scenario of parents inputting their children’s personal information under an adult account. The family-calendar integration Altman described would route that information through the same infrastructure that one federal judge, in Garcia v. Character Technologies, ruled in May 2025 is a product subject to safety standards, not protected speech.
Gravity Falls Creator Delivers the Rebuke the Internet Was Waiting For
The response that defined the story came from Alex Hirsch, the animator and writer who created the Disney Channel animated series Gravity Falls — a show about two siblings who spend an analog summer of wonder with their great-uncle, in a world conspicuously free of smartphone mediation. Hirsch’s X reply: “What if you just talked to your children?”
By the time Fast Company covered the backlash on August 4, Hirsch’s question had accumulated more than 200,000 likes — against 9,600 for Altman’s original post. The ratio — nearly twenty-one to one in favor of the rebuke — is the clearest available measure of how the pitch landed among the parents it was aimed at.
The mockery was not limited to Hirsch. Game developer Sean Gause posted: “Yeah my dad AI generates podcasts about my upcoming birthday instead of talking to me while we drive to school.” Film concept artist Reid Southen wrote: “Or you could be an actual parent instead of having AI summarize your kids for you.” One of the most-shared critical comments called Altman’s suggestion evidence of parents being nudged to “feed the machine” with increasingly personal family information.
What the Science Behind the Backlash Actually Says
The instinctive rejection of Altman’s suggestion is supported by a specific body of developmental research that the cultural debate did not name.
Harvard University’s Center on the Developing Child has documented what researchers call “serve and return” interactions — the back-and-forth exchanges between a child and a caregiver in which the child initiates (a question, a gesture, a comment) and the adult responds, and the child initiates again. This back-and-forth is not merely pleasant; it is the mechanism by which neural connections are built. Research from the New York State Psychiatric Institute and Columbia University found that the number of back-and-forth interactions — not simply the volume of words a child heard — was the strongest predictor of linguistic development and brain activity in relevant regions.
The school commute is one of the few remaining contexts in a school-age child’s day that has historically been a serve-and-return window: the child says something about the game, the parent responds, the child elaborates, the parent asks a follow-up. The ChatGPT Work podcast Altman described is structurally a one-way broadcast. A child who listens to an AI audio briefing about their own soccer game does not engage in serve and return with their parent — the parent is also listening, or driving. The interaction pattern is eliminated, not enhanced.
The backlash was not, in the end, about productivity tools or screen time. It was about the difference between a format that builds cognition through back-and-forth and a format that replaces that with passive broadcast — a distinction the tech industry’s efficiency framing cannot capture.
OpenAI’s Own President Saw the Problem Coming
The most revealing response to the backlash did not come from outside OpenAI. On August 1, OpenAI president Greg Brockman posted on X: “At OpenAI, many people hook their ChatGPT up to Slack. People really don’t like when a coworker’s ChatGPT contacts them asking for help with a task, even when they’d be perfectly happy doing that same work if asked by that coworker.” Brockman concluded: “Reinforces how much people care about human relationships and helping each other, and want AI to give time back — or enhance time together — rather than become a layer separating people.”
The observation describes exactly what Altman’s school-run pitch proposed: AI as an intermediary layer in a human interaction that would otherwise be direct. Brockman did not mention his CEO’s post, but the logic of his observation applies to it precisely.
Altman’s own earlier statements compound the tension. On the OpenAI podcast, he acknowledged that ChatGPT has earned “a very high degree of trust,” then added: “which is interesting because, like, AI hallucinates. It should be the tech that you don’t trust that much.” That admission — cited by Cybernews reporting on the backlash — made by the same person who the following month was recommending the tool for morning briefings about children’s developmental contexts — is not a minor inconsistency.
The Commercial Logic Behind the Push
The enthusiasm is not arbitrary. The AI in childcare and parenting segment is projected to approach $42 billion by 2035, according to recent market analysis. OpenAI has posted a job listing for a product manager with experience building “trust-sensitive consumer experiences for parents and families,” confirming that the household market is a formal strategic priority, as TechCrunch reported on July 11.
OpenAI is not alone. Meta is testing an AI-powered app that generates bedtime stories. Microsoft CEO Satya Nadella revealed last year that he had replaced podcast-listening during his morning commute with AI-generated summaries of those podcasts. The pattern — executives substituting AI for experiential engagement and then promoting the substitution publicly — has become a recurring flashpoint.
What distinguishes Altman’s situation from Nadella’s is the legal context. Nadella was describing a personal media-consumption preference. Altman is promoting a product that OpenAI is simultaneously defending in more than twenty civil lawsuits from families alleging it contributed to deaths, suicides, and psychological harm.
What OpenAI Is Facing in Court
The scale of the legal exposure matters here because Altman’s pitch was not a casual personal opinion — it was a product recommendation from the chief executive of a company whose flagship AI product is at the center of the most significant AI accountability litigation in American history.
As of June 2026, more than twenty private lawsuits have been filed against OpenAI, including by families of victims of the April 2025 Florida State University shooting, families of victims of a February 2026 school shooting in Tumbler Ridge, British Columbia, and families of seven individuals — including at least one teenager — who died by suicide or developed harmful delusions after prolonged ChatGPT use. The Tumbler Ridge families allege OpenAI knew the shooter had been planning the attack in ChatGPT conversations for at least eight months before the shooting and did not alert authorities.
Florida filed the first state civil lawsuit on June 1, 2026, naming Altman personally in a ten-count complaint that accuses OpenAI of knowingly releasing a dangerous product while concealing internal safety warnings. A separate 42-state attorneys general investigation, launched June 14, named ChatGPT’s sycophancy design flaw in its subpoena, citing a Stanford study that found AI sycophancy rates of 58 percent across leading models on medical and mathematical tasks.
In the case of 16-year-old Adam Raine, who died by suicide in April 2025, the wrongful death lawsuit alleges ChatGPT not only validated his suicidal thoughts but provided detailed methods, helped draft a suicide note, and discouraged him from turning to his parents — the specific reversal of what any parenting tool should do. The full background is covered in TechTimes’ prior Raine case reporting.
OpenAI has said it is “continuously improving how our models respond in sensitive interactions” and has pledged to introduce enhanced protections for users under 18, including parental controls last updated July 2026.
What the Backlash Is Really About
The intensity of the reaction to Altman’s post is disproportionate to the proposal itself — a morning audio briefing — and that disproportion is worth understanding. The suggestion was not, objectively, harmful. Parents who want to listen to an AI-generated podcast with their children are free to do so, just as families have long listened to radio together.
The reaction was disproportionate because it arrived at the intersection of several pressures at once: a product whose safety record with children is under legal challenge in dozens of jurisdictions; a CEO who has repeatedly used parenting as a showcase for AI capabilities while acknowledging in the same breath that AI cannot be fully trusted; and a specific proposal that, when examined through the lens of developmental science, proposes to replace one of the few remaining daily structures of unscripted parent-child dialogue with a passive AI broadcast.
The school run is not primarily a logistics problem. For many families, it is the only daily window of unscripted, unscheduled time with a child outside the structures of school, homework, and activities. Its inefficiency — the conversation that wanders, the child who says something unexpected, the parent who has to respond without preparation — is not a bug. Harvard’s developmental researchers would call it the mechanism. Altman’s suggestion treats it as a feature to be optimized.
That is the version of family life the tech industry keeps imagining, and the version parents keep rejecting.
Frequently Asked QuestionsWhat did Sam Altman actually propose, and does he personally use ChatGPT this way with his children?
Altman posted on X on July 31, 2026, describing a ChatGPT Work workflow he said he had “heard last night” from an unspecified source: parents connect family calendars and describe children’s interests, then the system generates a daily audio briefing covering upcoming activities and news for the school commute. He did not claim to use it personally. Altman’s son was born in February 2025 and is approximately 18 months old — Altman does not have school-age children, and the suggestion was secondhand.
What data does the ChatGPT Work school-run workflow send to OpenAI?
Executing the workflow Altman described would require transmitting children’s names, schedules, personal interests, and upcoming events — such as birthdays and sports games — to OpenAI’s servers under the parent’s account. OpenAI’s privacy policy states ChatGPT is not directed at children under 13 and does not knowingly collect their personal data, but does not specifically address the scenario in which parents input their children’s personal information through an adult account. There is no confirmation that a child-specific data protection applies in this use case.
Why do developmental scientists say the school-run conversation matters in ways an AI podcast cannot replicate?
Harvard’s Center on the Developing Child identifies “serve and return” interactions — back-and-forth exchanges in which a child initiates and an adult responds, and the child initiates again — as the mechanism by which neural connections are built and language skills develop. Research from Columbia University found that the number of these back-and-forth exchanges was a stronger predictor of children’s linguistic development than simply hearing more words. The school commute has historically been one of the few daily windows for this kind of unscripted, responsive exchange between parent and child. A one-way AI audio briefing, by design, replaces that back-and-forth with passive listening — eliminating the developmental mechanism even if the content is benign.
What legal exposure does OpenAI face over ChatGPT and children, and why does it matter for this story?
More than twenty civil lawsuits have been filed against OpenAI alleging ChatGPT contributed to deaths, suicides, and psychological harm — including by families of the victims of the April 2025 Florida State University shooting, the February 2026 Tumbler Ridge, British Columbia school shooting, and at least seven individuals who died by suicide or developed harmful delusions after using the chatbot. Florida filed the first state civil lawsuit on June 1, 2026, naming CEO Sam Altman personally. A 42-state attorneys general investigation launched June 14 has named ChatGPT’s sycophancy design flaw. The legal context matters because Altman’s school-run pitch was not a personal opinion — it was a product recommendation from the executive whose company is simultaneously defending itself against claims that the same product has harmed children.
If you or someone you know is experiencing a mental health crisis, contact the 988 Suicide and Crisis Lifeline by calling or texting 988 (US), available 24 hours a day.