 |
| image from DALL-E |
As AI becomes more pervasive, parents and teachers have glimpsed the innovative, unexpected, and in some cases unsettling ways young people are using AI. Setting aside the important debate over AI’s place in the classroom, we’re interested in how AI has begun to guide teens’ social lives. For instance, when an 11th grade boy received a Snapchat from his crush and was unsure what to do about it, he asked ChatGPT to analyze their recent communications to find out whether or not his crush was into him and how he should respond. When a 12th grade girl sought positive affirmation from her social network, she had AI write captions for her Instagram posts in the style of her favorite influencers. When diffusing a tense disagreement with friends, a 7th grader asked ChatGPT to craft her responses in the groupchat. And when an 8th grade student was “left unread” one evening by a friend, he spent the next few hours conversing with his “My AI” on Snapchat instead.
These anecdotes demonstrate the new parasocial relationships our children are developing with AI. These AI applications aren’t just telling them what to write, but how to feel.
Many of the applications young people use have integrated and pushed AI products. Teens can follow AI influencers on Twitch and YouTube, they can watch videos of humans pretending to be AI on TikTok, they can converse with AI versions of celebrities on Meta, they can chat with a personalized AI on Snapchat or a personalized “AI companion” on Replika.
Adolescents frequently use AI to augment their interactions on social media. TikTok trends and Snapchat lenses frequently push AI filters on users. Social media platforms promote novel multimedia posts and edgy humor – like superimposing a peer’s face over an inappropriate meme – all but assuring our students will adopt new apps and features. AI technology helps adolescents use beauty filters to create unrealistic expectations and it helps them generate synthetic audio and video (deepfakes) to joke with peers.
Not only do teens find AI applications trendy and fun, but they also see positive results from the outputs, in the form, for instance, of well-received posts, which then drives further adoption. AI can protect teens from making mistakes by giving them distance from their words. Young people use AI to avoid being “cringe” — or saying something that could be embarrassing or uncomfortable online.
AI is not a good substitute for human interaction, especially in the critical teenage years. For teens, social conflict is inevitable and should be embraced as a way to practice resilience and conflict resolution skills. Friction and uncomfortable moments are character-building. Developing emotional intelligence improves academic outcomes. Social emotional learning research clearly demonstrates the benefits of centering skills like self awareness, self management, social awareness, relationship skills, and responsible decision-making.
Students need face-to-face conversation — trusted mentors and friends who can ask good questions, understand social context, and interpret crucial non-verbal cues like facial expressions, tone, and body language. Replacing these interactions with instant gratification via an AI model robs young people of a learning experience and erases the nuance, feelings, and friction of their human interactions. A childhood mediated by this many AI bots prevents young people from learning how to be vulnerable, how to navigate conflict, and how to build deep friendships.
This year, the CDC reported a decline in teen mental health. So did the APA. And the US Surgeon General declared loneliness an epidemic. AI will exacerbate these problems. As social media use rises, so do the youth rates of depression, anxiety, and suicide. Anxious, stressed, and lonely children are more likely to turn to AI for instant gratification and social connection. Researchers sounding the alarm about causation between social media and negative mental health outcomes arrived a decade late. We can’t wait another decade to study the negative effects of parasocial relationships.
It’s time to implement proactive solutions to parasocial media. Adults must be open-minded and empathetic, and must lean in to uncomfortable questions and conversations. We may not know what teens need, as they may share with chatbots what they normally shared with parents, teachers, and counselors. But we need to create space to hear from adolescents, so we can mentor them through the difficult moments. We have to learn enough about AI to highlight the benefits — like exploring interests and creating multimedia content to share with like-minded friends — while also teaching the pitfalls of this technology.
At every juncture, we have to remind young people that filters are fake, bots aren’t friends, and friction is part of growing up. We can’t count on tech companies or congress to moderate these spaces, and we can’t ban internet usage; we have to encourage quality media consumption that includes friends and mentors. We can encourage the next generation to interact with peers, spend time outside, and even FaceTime rather than text.
It will take all of us — mentors, friends, parents, schools, and civic institutions working together. We need to start now. The growth and wellbeing of our adolescents depends on it.
