Jonathan Haidt, the NYU Stern School of Business professor whose book "The Anxious Generation" ignited worldwide discussion on smartphones and social media effects on youth mental health, now raises alarms about AI toys aimed at children.

At a recent TED Talk, Haidt noted that the expanding sector of AI-driven toys and chatbots might disrupt kids' connections to parents by acting as primary emotional supports.

"We are witnessing rapid growth in the AI toy sector, with chatbots integrated into dolls and teddy bears," Haidt stated.

Although he avoided naming particular items, a new wave of AI-enhanced toys and companions has appeared in recent years.

Instances include Moxie, an AI social robot intended as a friend, tutor, and guide for children, along with chatbot dolls and bears from firms such as Curio. Major toy maker Mattel has also teamed with OpenAI to develop AI products, highlighting the swift pace of change in this area.

Market research firm Grand View Research projects the worldwide smart toy market, encompassing AI toys, will expand from $14.39 billion in 2025 to $44 billion by 2033.

The social psychologist explained that these items are built to respond instantly to children, providing comfort, dialogue, and company on demand.

Such constant presence may heighten their appeal for young users.

"These chatbots respond instantly to the child and remain available for support when needed, while parents are frequently occupied," Haidt said.

The Peril of Artificial Bonds

Haidt's caution arrives amid AI companions already altering social and romantic interactions for many people.

Users report developing attachments to AI partners, depending on chatbots for regular company, and worrying about losing those virtual ties.

Haidt focuses specifically on how children build emotional connections.

Young children naturally seek out consistent responders to their needs, he noted. Should AI companions prove more reliably available than parents, they might assume an unhealthy position in a child's emotional growth.

"If the chatbot stays highly responsive while parents are less so, the child's attachment system may latch onto the chatbot instead, which could harm their bond with parents," Haidt said.

A Voice for Caution on Technology

The alert fits into Haidt's larger case for what he terms techno-skepticism.

He argued that society advanced too fast in letting social media firms influence children's social worlds and educational tech firms alter classrooms.

"Social media firms assumed control over kids' social lives, damaging their interactions and mental health," he said. "EdTech firms took over schools, and the results seem more harmful than helpful."

Now, he added, AI firms target an even closer aspect of childhood: personal relationships.

"AI companies seek to become children's friends and therapists," Haidt said. "What might go wrong?"

Although AI therapists could eventually play a part, he stressed that firms must not release emotionally influential AI items for children without extensive safety evaluations.

"Provide nothing that suggests understanding or caring for the child," Haidt said. "Because it does not."