At 3 a.m., with a newborn crying and sleep out of reach, most parents reach for their phones.
Sam Altman, CEO of OpenAI and a new father, did the same. But instead of searching Google, he opened ChatGPT.
On OpenAI’s latest podcast, Altman revealed that he regularly used the chatbot during the early days of his son’s life.
He asked the chatbot questions about baby behavior, development, and sleep routines. Now that his child is three months old, Altman continues to consult ChatGPT for parenting advice.
“I don’t know how I would’ve done that,” he said. “Clearly, people have raised children without ChatGPT. But it helped a lot.”
The Convenience of ChatGPT
Using AI for parenting questions is not entirely new. Many parents already use online search engines, apps, and forums.
ChatGPT, however, provides fast, conversational responses. It mimics a knowledgeable assistant rather than delivering a list of links.
Still, ChatGPT is not always accurate. It can generate “hallucinations”, false or misleading answers stated with confidence.
This raises a major issue: should parents rely on AI for decisions that affect a child’s health or well-being?
Even Altman acknowledges the limitations. He understands the risks and believes the public must remain cautious.
A Familiar Pattern
Altman’s experience echoes what many modern parents face. The internet has long been a source of support, and anxiety, for those raising children.
From all over the internet, advice comes quickly and often contradicts itself. ChatGPT, however, is a cleaner, more direct experience.
It eliminates the need to sift through long forum threads or conflicting opinions. Yet it remains an AI tool, not a pediatrician or parenting expert.
In moments of panic, some may accept any response that sounds reasonable. That is where the danger lies.
Children and Chatbots
The podcast also explored another topic: children using AI themselves. Altman and former OpenAI science communicator Andrew Mayne discussed a story about a parent who let ChatGPT’s voice mode speak with his toddler.
The child, obsessed with Thomas the Tank Engine, talked with the chatbot for over an hour. “Kids love voice mode,” Altman said.
This scenario is a reflection of AI becoming part of everyday life, even for very young users. But ChatGPT is not designed for children.
Its policies state that it should not be used by anyone under 13 without supervision. It lacks parental controls.
It also does not follow child-specific content guidelines. Altman recognizes this. He noted that society will need to build new guardrails as AI becomes more integrated into family life.
“There will be problems,” he said. “People will develop problematic parasocial relationships. We will need to address that.”
Also read: The New Trend of AI Dark Cartoons
The AI Companion
This situation resembles earlier concerns around screen time and “iPad kids.” Parents have long debated how much tech is too much.
With AI, the conversation shifts. Instead of just watching a screen, children can now interact with it.
Unlike videos or games created by professionals, AI responses are generated in real time. They are not reviewed by child development experts.
And they change depending on the user’s input. This creates new risks and new responsibilities.
Parents must now decide not only how much screen time is appropriate, but also how much AI interaction is safe.
