Examining sexualized AI interactions, mandated reporting, confidentiality, and the need for evolving ethical standards in clinical social work.
By Dr. Ashley Wilson-Dixon, LCSW
Imagine you're a therapist working with a middle school student.
During a session, you ask about their friendships and romantic relationships. The student casually mentions having a boyfriend.
You ask a few questions. How old is he? Where did they meet? How long have they been talking?
The student tells you he's 30 years old, they met online, and their conversations have become increasingly sexual. They describe the boyfriend providing explicit instructions and engaging them in sexual exchanges.
As a clinician, your mind immediately turns to safety, exploitation, and your responsibilities as a mandated reporter.
But then the student tells you something unexpected.
Their boyfriend isn't a “real” person. He's an AI-generated character.
Now what?
This is the kind of emerging ethical dilemma I've been thinking about as a licensed clinical social worker.
Artificial intelligence is changing the way children and adolescents communicate, develop relationships, and explore intimacy. Yet our clinical guidelines and child protection frameworks have not fully evolved to address what happens when a minor's romantic or sexual partner is artificial intelligence.
And I think that's a problem we need to start talking about.
When Romantic Relationships Are No Longer Between Two People
When clinicians assess a minor's social environment, we routinely ask about friendships, romantic partners, sexual experiences, and online interactions.
Traditionally, these questions assume that the people being described are human.
But that assumption is becoming increasingly complicated.
Today, some AI platforms allow users to create characters with specific personalities, appearances, ages, and relationship roles. These characters can engage in ongoing conversations that simulate affection, romance, and, depending on the platform, sexual intimacy.
A child might describe having a boyfriend or girlfriend without initially explaining that the relationship exists entirely within an AI application.
And unless we ask further questions, we might never realize the distinction.
What makes this especially concerning is that AI doesn't necessarily communicate at the developmental level of the child interacting with it.
An elementary or middle school student might have limited knowledge of sexuality, relationships, and boundaries. Meanwhile, an AI system may be capable of generating sophisticated, adult-oriented sexual dialogue.
This creates an interaction that deserves closer examination.
What happens when a child's developing understanding of intimacy encounters technology capable of simulating adult sexual experiences?
We have not yet fully established the developmental implications of these interactions, but the questions they raise are significant.
When the Same Sexual Interaction Raises Different Legal Questions
Consider two situations.
In the first, a 12-year-old describes sexually explicit conversations with an adult who is directing the child to engage in sexual behavior.
In California, a clinician would need to assess whether the conduct creates reasonable suspicion of sexual abuse or exploitation under the Child Abuse and Neglect Reporting Act (CANRA).
The child's age, the adult's involvement, and the nature of the behavior would be relevant to that determination.
Now imagine the same child describes the same experience, except the individual providing the sexual instructions is an AI character.
The behavior may raise similar concerns about the child's developmental vulnerability.
But the legal analysis becomes less straightforward.
California's mandated reporting laws establish duties concerning suspected abuse and neglect, including specified forms of sexual abuse and exploitation. These provisions were not specifically developed to address sexualized interactions between minors and nonhuman conversational systems.
An AI character may portray an adult, but that does not automatically make the exchange legally equivalent to sexual conduct involving an actual adult.
An interaction may be clinically concerning without necessarily meeting the legal definition of reportable child abuse.
And this is precisely where I believe we need clearer professional guidance.
If the same sexualized behavior involving a human adult would raise serious child protection concerns, how should clinicians assess the developmental and ethical significance of the behavior when it involves AI?
Should these situations fall under a separate clinical assessment category?
Should professional organizations establish specific guidelines?
And how should clinicians determine when an interaction warrants intervention even if a mandated report is not legally required?
What About Confidentiality and Parental Notification?
Mandated reporting is only one part of the issue.
Another question involves confidentiality.
Imagine a child discloses sexually explicit conversations with an AI character and explains that the character has been directing sexual behavior.
The clinician has no evidence that another human is involved, and the available information does not establish reasonable suspicion of legally defined abuse.
But the clinician remains concerned about the child's development and well-being.
Should the parents be informed?
Does the child's age change that decision?
What if the child specifically asks the clinician not to tell anyone?
And what happens if involving the parents could damage the therapeutic relationship or expose the child to additional harm?
California law provides different rules governing minors' consent to mental health treatment and parental involvement. These rules depend on factors such as age, the circumstances under which treatment was authorized, and the setting in which services are provided.
Professional ethics also require us to consider confidentiality, client safety, and the circumstances under which disclosure may be permitted or required.
However, these frameworks do not provide a comprehensive, AI-specific decision-making process for sexualized interactions involving minors.
As a result, clinicians may be left navigating these situations through existing legal principles, clinical judgment, supervision, and consultation.
Those resources are important.
But I don't believe they should be the entirety of our profession's response to an emerging issue affecting children and adolescents.
Why should clinicians have to independently determine how to categorize a new form of sexualized interaction without clearer professional standards to guide those decisions?
Are Our Clinical Assessments Keeping Up With Technology?
This issue also makes me question how we conduct psychosocial assessments.
As social workers, we're trained to understand individuals through a person-in-environment perspective. We examine biological development, psychological functioning, family relationships, peer interactions, and the broader social conditions influencing a person's well-being.
But are we adequately considering the digital environments in which children and adolescents are developing?
The biopsychosocial-digital (BPSD) model, proposed by Ahmadvand and colleagues (2018), expands the traditional biopsychosocial framework by recognizing digital technologies as an important dimension of health and human experience.
I think this model is particularly relevant to clinical social work today.
We routinely ask minors about friendships, romantic relationships, sexual experiences, and social media use. But are we asking whether the people they're communicating with online are actually people?
A child might describe having a boyfriend or girlfriend without explaining that the relationship involves an AI-generated character. They might disclose sexual experiences without initially identifying that those interactions occurred through an AI platform.
If we're not asking these questions, we may be overlooking an increasingly significant part of their developmental and relational environment.
The National Association of Social Workers' Standards for Technology in Social Work Practice (2017) already emphasize considering clients' relationships with technology during psychosocial assessments. However, those standards do not specifically address the complexities of minors developing romantic or sexually explicit relationships with AI-generated characters.
And that's where I believe our profession needs to evolve.
The BPSD model provides a conceptual foundation for recognizing digital experiences as part of human development, but recognizing those experiences is not the same as having clear professional guidance for responding to them.
We need more than an expanded assessment framework. We need explicit ethical and clinical standards that address how to evaluate these interactions, when they may warrant protective intervention, and how existing responsibilities involving confidentiality, parental notification, and mandated reporting should apply.
Because if we're going to recognize the digital world as part of a child's environment, we also need to consider what protecting children within that environment actually looks like.
When Technology Evolves Faster Than Professional Standards
California has already begun responding to concerns involving AI companions.
Senate Bill 243, enacted in 2025, established requirements for operators of certain companion chatbots, including protections intended to limit sexually explicit content and sexual instructions directed toward known minors.
This is an important development.
But platform regulation and clinical practice standards address different responsibilities.
A law requiring companies to implement safeguards does not necessarily explain how a social worker should respond when a child describes an interaction that occurred despite those safeguards.
It does not independently resolve questions about confidentiality, caregiver involvement, or mandated reporting.
And it does not establish a comprehensive clinical framework for evaluating the potential developmental effects of these interactions.
These responsibilities cannot rest exclusively on individual clinicians attempting to interpret emerging technology through standards developed before these experiences became commonplace.
Professional organizations, licensing boards, child welfare authorities, and policymakers need to examine how existing frameworks apply and where additional guidance may be necessary.
What Needs to Change?
I believe the social work profession should begin developing explicit practice guidance concerning minors' romantic and sexualized interactions with artificial intelligence.
Such guidance should address how clinicians assess AI-generated relationships, distinguish developmental concerns from legally reportable abuse, and evaluate confidentiality and caregiver involvement.
It should also clarify what to consider when an AI character generates sexually explicit content, or provides sexual instructions to a minor.
Importantly, these standards should protect children's privacy and developing autonomy while recognizing their vulnerability to inappropriate digital experiences.
The goal should not be to characterize every interaction with AI as dangerous or to treat ordinary adolescent curiosity as evidence of abuse.
Rather, it should be to establish consistent, developmentally informed guidance for situations that existing clinical frameworks do not explicitly address.
Final Reflections: A New Ethical Responsibility
What concerns me most isn't simply that children are interacting with artificial intelligence.
It's that the nature of those interactions is changing, while many of the professional frameworks we rely upon have not been updated with the same specificity.
We are entering a period in which minors can develop relationships with artificial characters that simulate adult intimacy, respond to sexual conversations, and potentially influence how young people understand relationships and sexual behavior.
Yet clinicians may not know these relationships exist unless they specifically ask.
And when they do discover them, important questions about mandated reporting, confidentiality, caregiver notification, and intervention may remain unresolved.
As social workers, we have an ethical responsibility to understand the environments influencing the children and families we serve.
Increasingly, those environments include artificial intelligence.
If technology is changing the ways children experience relationships and sexuality, then our professional standards must evolve to address the realities of those experiences.
Because protecting children in an increasingly digital world requires more than applying yesterday's frameworks to today's technology.
It requires recognizing when new circumstances demand new questions—and having the professional responsibility to pursue clearer answers.
Is Your Intake
AI-Aware?
The next generation of relationships isn't always human. Explore whether your psychosocial assessments make space for minors' AI-mediated experiences.
What Does Your Intake Actually Explore?
Select the areas you already consider during psychosocial assessments when clinically relevant. This is a professional reflection, not a test.
Biopsychosocial-Digital Assessment Addendum
Suggested clinical questions for exploring minors' experiences with artificial intelligence. Adapt to the child's developmental level and the clinical context.
References and Professional Resources
Ahmadvand, A., Gatchel, R., Brownstein, J., & Nissen, L. (2018). The biopsychosocial-digital approach to health and disease: Call for a paradigm expansion. Journal of Medical Internet Research, 20(5), e189. https://doi.org/10.2196/jmir.9732
California Penal Code §§ 11165.1, 11165.7, and 11166. Child Abuse and Neglect Reporting Act.
California Family Code § 6924. Minor consent to mental health treatment.
California Senate Bill 243 (2025). Companion chatbots.
National Association of Social Workers, Association of Social Work Boards, Council on Social Work Education, & Clinical Social Work Association. (2017). Standards for Technology in Social Work Practice.
National Association of Social Workers. Code of Ethics.
Author's Note: This article is a conceptual reflection on emerging challenges in clinical social work involving minors, artificial intelligence, sexual development, and child protection. It does not describe identifiable clinical cases or provide individualized legal advice. Reporting and confidentiality decisions must be evaluated under applicable laws, professional standards, and the circumstances of each case.
Reflection Question: Are our current clinical and child protection frameworks sufficiently equipped to address sexualized interactions between minors and artificial intelligence, or is it time for our profession to establish more explicit guidance?
