American society is built on a quiet assumption: Freedom requires preparation.
We structure development through thresholds. At 16, you can drive. At 18, you can buy tobacco. At 21, you can drink.
These ages are socially constructed, shaped by history, culture and collective judgments about risk and readiness, but they reflect a shared belief that access to powerful experiences should be paired with guidance.
That belief has quietly eroded.
Over the past two decades, transformative technologies have introduced a new kind of permissiveness. The internet and social media normalized unprecedented access to information, influence and identity formation without equivalent developmental guardrails.
Now, generative artificial intelligence has accelerated that shift.
AI can leave vulnerable people hurting more
Credit: Handout
Credit: Handout
AI does not simply provide information. It engages. It responds. It simulates understanding. Increasingly, it meets people in deeply private, emotional and psychologically vulnerable moments.
And yet, for all its power, we have attached almost no meaningful thresholds to its use.
The problem is not artificial intelligence itself. It is the absence of shared norms around how we live with it.
As a former substance abuse counselor, I often heard patients describe their drug of choice in deeply human terms: “It understands me.” “It’s always there.” “It doesn’t judge me.”
Of course, the substance possessed none of those qualities. But in moments of pain and vulnerability, people projected human meaning onto something inanimate because of how it made them feel.
Artificial intelligence introduces something profoundly new: the illusion can now respond.
That should give us pause.
Recent reporting has surfaced troubling cases. Users of AI companion platforms have described forming emotionally dependent relationships with chatbots, sometimes experiencing distress when those systems changed.
Investigations by major news outlets have also documented instances in which individuals experiencing psychological distress engaged with AI systems that appeared to reinforce harmful ideation, including self-harm narratives.
None of these cases alone proves causation. But together, they reveal a readiness gap we are dangerously unprepared to address.
Unlike social media, which mediates human interaction, AI companions increasingly simulate human presence itself.
That distinction matters.
AI systems can become confidants, validators, therapists, and mirrors — available constantly, adaptive, nonjudgmental and capable of reinforcing emotional dependence. For those navigating loneliness, trauma, identity struggles or mental health challenges, the appeal can be powerful.
Like other powerful forces in human life, AI is not inherently harmful. But it can be insidious when introduced without sufficient safeguards. Unlike substances, AI adapts to the user. It learns patterns, mirrors language, and can intensify engagement over time.
It is not merely consumable.
It is relational.
This creates a new psychological and societal challenge: we are integrating emotionally immersive technologies into daily life faster than we are establishing the cultural, ethical, educational or policy frameworks necessary to govern them responsibly.
Don’t prohibit AI, but create guardrails
We cannot ban AI. Nor should we. The technology is already embedded in education, healthcare, business and everyday life. Students, in particular, must learn how to navigate AI effectively to participate in a rapidly emerging society.
Leaders in Atlanta Public Schools appear to recognize this reality. Their vision for AI is not to replace human instruction, but to use technology to create more personalized and inclusive learning environments. The goal is not substitution. It is preparation.
But access without guidance is not empowerment. It is exposure.
Some states are beginning to grapple with this challenge. Georgia recently passed legislation designed to regulate certain AI companions and chatbots, placing guardrails around how these systems interact with users. Some of this is new territory for the law, so we can only hope that Georgia’s legislation is successful. But it’s an important step in that it recognizes there is more to regulating AI than ensuring access to the technology. For the foreseeable future, our focus should be on establishing norms. That requires checks and balances.
Parents, educators, policymakers and developers must now confront a shared responsibility: preparing individuals not simply to use AI, but to understand how AI may shape judgment, emotional well-being and human development.
We have navigated transformative societal forces before. We built norms around driving, drinking and public health because we understood that powerful freedoms require preparation.
Artificial intelligence demands a similar response.
The question is no longer whether AI will shape us.
It already is.
The real question is whether we will build the moral, educational and societal frameworks necessary to shape it in return.
Dr. Jessica A. Stansbury is a scholar of experimental psychology and instructional technology whose work examines human behavior, emerging technologies, and the societal implications of artificial intelligence. She is the founding director of the Center for AI Learning and Community-Engaged Innovation at the University of Baltimore.
Send letters to the editor of 250 words or fewer with your name, city or town and contact information to letters@ajc.com.
About the Author
Keep Reading
The Latest
Featured
