The Rise of AI as an Emotional Proxy

The article, penned by psychologist and author Jonathan Haidt, highlights how Gen Z and younger millennials are leveraging AI to outsource the “hardest part of growing up”—developing resilience through failure, rejection, and unfiltered human interaction. Haidt argues that while these tools offer immediate relief, they erode essential skills like emotional regulation and interpersonal negotiation.

“Young people are using AI to simulate conversations they fear having in real life,” Haidt writes. Examples abound: teens prompting AI to generate scripts for confronting parents about curfews, apologizing to friends after arguments, or even ending romantic relationships. Platforms like Replika and Pi.ai have seen explosive growth, with millions of downloads among users under 25, per recent App Annie data.

Expert Warnings and Real-World Impacts

Psychologists echo Haidt’s concerns. Dr. Jean Twenge, co-author of iGen, notes in a recent interview that over-reliance on AI could exacerbate the mental health crisis already gripping youth. “We’re seeing skyrocketing anxiety rates—up 50% since 2010 per CDC data—partly because kids aren’t practicing discomfort tolerance,” she said.

Anecdotal evidence supports this. In a viral TikTok trend, users share “AI vs. Real Life” videos, juxtaposing smooth AI-handled interactions with their bungled real attempts. One 19-year-old college student told The Guardian, “I let ChatGPT write my apology to my roommate. It worked great—no yelling. But now I don’t know how to do it myself.”

Tech Industry’s Response and Counterarguments

AI developers defend their tools as supplements, not replacements. Character.AI’s founders stated in a blog post, “Our platform empowers users to practice social skills in a safe space, building confidence for real-world application.” Data from the company shows 70% of users under 18 report improved communication after sessions.

Yet critics like Haidt counter that simulated practice lacks stakes. “AI doesn’t get hurt, doesn’t judge, doesn’t remember,” he writes. A 2025 study from Stanford’s Human-Centered AI Institute found that participants who used AI for emotional coaching scored 25% lower on empathy tests compared to those practicing with peers.

Broadening the Phenomenon: Beyond Personal Drama

The trend extends to professional spheres. Job seekers use AI to rehearse interviews, with tools like Interviewing.io generating hyper-personalized feedback. High schoolers prompt AI for college essays, risking authenticity penalties amid rising plagiarism detection. Even therapy is disrupted: 40% of BetterHelp users under 30 admit supplementing sessions with AI chatbots, per internal surveys leaked last year.

Societal Implications and Calls for Balance

As AI permeates education and social media, educators are adapting. Schools in California and New York have introduced “AI Literacy” curricula, teaching students to recognize when tech undermines growth. Parents’ groups advocate for age gates on emotional AI apps, citing EU regulations like the Digital Services Act as models.

Haidt proposes a radical fix: mandatory “unplugged” periods where youth tackle challenges sans AI. “Growth happens in the messiness,” he asserts. While not all agree—some view AI as a democratizing force for introverts—the debate underscores a pivotal question: In automating empathy, are we automating away maturity?

With AI advancing rapidly—OpenAI’s latest model now handles nuanced emotional role-play—the tension between convenience and character-building intensifies. As one 16-year-old put it on Reddit: “AI makes life easier, but am I even living it?”