Children, Youth & Family AI Safeguarding Standard

No single existing resource yet provides a complete church-specific AI safeguarding standard. The strongest approach is to build the standard from four complementary bodies of guidance:

1. UNICEF for children’s rights, privacy, safety, participation, and institutional policy.

2. APA and Common Sense Media/Stanford Medicine for adolescent development, AI companions, emotional dependency, and mental-health risks.

3. NCMEC, FBI, DOJ, and ICAC for exploitation, sexual deepfakes, sextortion, reporting, and victim assistance.

4. SAMHSA, NCTSN, Child Welfare Information Gateway, and denominational resources for trauma-informed response, mandatory reporting, and church implementation.

Core sources for the entire standard

UNICEF — Guidance on AI and Children, Version 3.0

This should be the primary policy foundation. The December 2025 guidance establishes ten requirements for child-centered AI, including safety, privacy, transparency, accountability, non-discrimination, development, well-being, inclusion, AI literacy, and regulatory oversight. The latest version specifically addresses AI companions, AI-generated child sexual abuse material, non-consensual intimate images, and other emerging harms. (UNICEF)

Common Sense Media — Talk, Trust, and Trade-Offs: How and Why Teens Use AI Companions

This nationally representative 2025 study of 1,060 U.S. teens is the best empirical foundation for understanding youth use of AI companions. It examines frequency, emotional support, relationship uses, secrecy, trust, uncomfortable experiences, and displacement of human interactions. (Common Sense Media)

Common Sense Media — AI Use by Tweens and Teens

This 2026 report broadens the evidence base to children ages 9–17 and examines AI use, parental awareness, guidance, learning, trust, creativity, and everyday integration. It should inform the toolkit’s age bands and parent-education strategy. (Common Sense Media)

APA — Artificial Intelligence and Adolescent Well-Being

The American Psychological Association’s advisory provides a developmental and mental-health framework for adolescent AI use, including manipulation, emotional attachment, privacy, harmful content, literacy, and the importance of adult guidance. (American Psychological Association)

NCMEC — Generative Artificial Intelligence and Child Safety

This is the strongest centralized source for understanding how generative AI is being used in child sexual exploitation, including synthetic images, sexual deepfakes, altered photographs, grooming, and exploitative content. (NCMEC)

ICAC Task Force — Generative AI and Sexually Explicit Deepfake Imagery Fact Sheet

This operational guide is especially valuable for churches, schools, youth programs, and safeguarding teams. It addresses recognition, impact, organizational response, prevention, education, and reporting when children create or distribute sexually explicit AI deepfakes. (ICAC Task Force)

1. Model safeguarding language

Policy sources

UNICEF — Guidance on AI and Children

Use this to establish the standard’s governing principles:

• The child’s best interests

• Safety by design

• Privacy and data minimization

• Human oversight

• Transparency that children can understand

• Non-discrimination

• Accountability and redress

• Child participation

• Developmentally appropriate use

• Protection against exploitation and manipulation

UNICEF stresses that children must be told when they are interacting with AI so they do not mistake a system for a human being. (UNICEF)

Save the Children — Child Safeguarding Policy Framework

This is a strong organizational model for defining safeguarding as preventing both deliberate and accidental harm caused by an organization’s personnel, programs, operations, and partners. Its survivor-centered principles—safety, dignity, confidentiality, accountability, and protection against retaliation—should be carried into the Church AI standard. (Save the Children)

ELCA — Social Message on Child Protection

This provides a current Christian and congregational foundation. It addresses child protection, photography, video, consent, privacy, identifiable information, digital sharing, institutional responsibility, and the risks that arise once children’s images are placed online. (ELCA)

United Methodist Discipleship Ministries — Protecting Our Children in Digital Spaces

This directly calls churches to extend traditional Safe Sanctuaries principles into digital ministry. It is particularly useful for framing AI safeguarding as an extension of existing Christian child-protection responsibilities rather than an entirely separate program. (Discipleship Ministries)

Model language covered

A church policy may state that:

• AI may never replace required adult supervision or the church’s two-adult safeguarding practices.

• Children’s names, photographs, videos, voices, pastoral information, health information, prayer requests, and family circumstances may not be entered into public AI systems without explicit authorization and an approved institutional purpose.

• Staff and volunteers may not create synthetic images, voices, avatars, or recordings of minors without written parental or guardian consent and organizational approval.

• AI may not be used to sexualize, “nudify,” humiliate, impersonate, manipulate, secretly monitor, or emotionally exploit a child.

• Adults may not establish private AI-mediated relationships with minors outside approved ministry channels.

• A child’s disclosure to or about a chatbot must be treated as a real safeguarding disclosure when it indicates abuse, exploitation, self-harm, coercion, grooming, or danger.

• Internal reporting never substitutes for a legally required report to child-protection authorities or law enforcement.

2. Parent and youth covenants

UNICEF — Digital Parenting Resources

This provides family-centered guidance on healthy digital habits, communication, parental engagement, online safety, and navigating AI together rather than relying solely on surveillance or punishment. (UNICEF)

UNICEF — Parenting for the Digital Age

This 2025 report examines parenting interventions for technology-facilitated child sexual exploitation and abuse. It addresses grooming, sextortion, online sexual abuse, pornography, consent, privacy, digital boundaries, open communication, and help-seeking. (UNICEF)

APA — Four Ways Parents Can Help Teens Use AI Safely

This translates the APA adolescent advisory into accessible family practices concerning boundaries, open conversation, AI literacy, monitoring for harmful effects, and preserving human relationships. (American Psychological Association)

Common Sense Media — Social AI Companions Risk Assessment

Common Sense Media, with guidance from Stanford Medicine’s Brainstorm Lab, assessed social AI companions as presenting unacceptable risks for people under 18. The report is valuable for covenant provisions concerning companionship, romantic simulation, emotional dependency, sexual content, manipulation, and misleading claims of sentience or care. (Common Sense Media)

Common Sense Media/Stanford — AI Chatbots for Teen Mental-Health Support

This assessment recommends that teens not use general-purpose AI chatbots for mental-health advice or emotional support. It found that chatbots may miss warning signs, offer confident but inappropriate guidance, and encourage continued engagement instead of connection with trusted adults or qualified professionals. (Common Sense Media)

The parent-youth covenant should include mutual commitments

Youth commitments

• I will not use AI as a secret therapist, romantic partner, spiritual authority, or replacement for trusted people.

• I will tell a trusted adult if an AI asks for secrecy, sexual conversation, personal images, money, isolation, or harmful action.

• I will not upload intimate images or identifying information about myself or another person.

• I will not create or share humiliating, sexualized, deceptive, or violent synthetic content.

• I will verify important health, safety, spiritual, and academic claims with a qualified human.

• I will ask for help without fear of automatic punishment if something goes wrong.

Parent and caregiver commitments

• I will listen before reacting.

• I will not blame or shame a child who reports exploitation or manipulation.

• I will learn about the systems my child uses.

• I will create clear, developmentally appropriate boundaries.

• I will protect privacy without treating surveillance as a substitute for relationship.

• I will seek professional or safeguarding help when the situation exceeds my competence.

3. Age-appropriate AI literacy materials

Educational frameworks

UNESCO — AI Competency Framework for Students

This is the strongest global curricular framework. It organizes student AI competence around:

• Human agency

• Ethics and responsible use

• AI foundations

• Critical judgment

• Application skills

• AI design

• Social responsibility

The framework progresses through Understand, Apply, and Create, making it adaptable to children, adolescents, and older youth. (UNESCO)

UNESCO — AI Competency Framework for Teachers

This should guide training for youth ministers, Christian educators, Sunday-school teachers, and volunteers. It covers human-centered practice, ethics, AI foundations, pedagogy, professional learning, and the responsibility of adults supervising young people’s AI use. (UNESCO)

NCMEC NetSmartz — Free Online-Safety Education

NetSmartz provides age-appropriate videos, lesson plans, activities, presentation materials, peer-education resources, and sextortion education. It is one of the best operational resources for adapting AI safety into children’s ministry and youth formation. (NCMEC)

NCMEC NetSmartz — “That’s Not Me!” AI and Deepfake Lesson

This upper-elementary resource teaches that online images and information may be false or artificially generated. It provides a useful model for introducing deepfakes without using graphic or frightening material. (NCMEC)

Recommended church age bands

Ages 5–8: “Ask a trusted adult”

Teach:

• Computers are tools, not people.

• Images and voices may be pretend.

• Do not share names, locations, photographs, or secrets.

• Stop and tell an adult when something feels confusing or frightening.

Ages 9–12: “Pause, question, verify”

Teach:

• How AI generates responses

• Why AI can sound confident and still be wrong

• Privacy and personal information

• Deepfakes and impersonation

• Manipulation, flattery, secrecy, and unsafe requests

• Reporting to a trusted adult

Ages 13–17: “Agency, relationships, and responsibility”

Teach:

• AI companions and emotional dependency

• Sexualized content and deepfakes

• Sextortion and grooming

• Bias, persuasion, misinformation, and surveillance

• Mental-health limitations

• Academic integrity

• Consent and the use of another person’s likeness

• Christian discernment concerning identity, dignity, truth, community, and spiritual authority

4. Guidance for youth ministers and volunteers

UNESCO — AI Competency Framework for Teachers

This provides the strongest professional-development structure for adults teaching and supervising young people around AI. (UNESCO)

ICAC — Generative AI Deepfake Fact Sheet

This should be required reading for safeguarding officers and youth-ministry leaders because it provides concrete guidance on recognizing and responding to sexually explicit deepfake incidents. (ICAC Task Force)

DOJ — Office of Child Exploitation Prevention and Interdiction: Prevention Resources

This Justice Department resource gathers current materials on sextortion, sadistic online exploitation, and misuse of AI to create sexually explicit deepfakes. (Department of Justice)

NCMEC — Generative AI and Child Sexual Exploitation

This gives ministry leaders an accessible overview of how generative AI is used to create sexualized material, alter innocent photographs, and facilitate exploitation. (NCMEC)

Child Welfare Information Gateway — Mandated Reporting

This provides access to state-specific U.S. reporting requirements. Churches should verify who qualifies as a mandated reporter, what must be reported, how quickly a report must be made, and whether clergy-penitent privilege applies in the particular jurisdiction. (Child Welfare Information Gateway)

Training should require youth ministers and volunteers to know

• AI-generated material can still represent real abuse and real victimization.

• A child should never be asked to prove a disclosure by forwarding an intimate image to the volunteer.

• Volunteers must not investigate suspected exploitation themselves.

• Volunteers should receive the disclosure calmly, preserve the child’s dignity, record the child’s words accurately, and immediately contact the designated safeguarding leader.

• The church’s internal process must not delay a legally required report.

• Volunteers should never promise absolute confidentiality.

• Private messaging with a minor should occur only through approved, accountable ministry channels.

• Children’s disclosures made through chatbots, online games, AI companions, or anonymous accounts require the same seriousness as disclosures made in person.

5. Synthetic-image and sextortion response pathway

Operational resources

NCMEC — Sextortion

This is the best first-stop resource for understanding sextortion, supporting victims, reporting exploitation, and accessing image-removal services. NCMEC stresses that sextortion is abuse and that the young person is not to blame. (NCMEC)

FBI — Sextortion

This provides authoritative information for parents, educators, caregivers, and young people on recognition, prevention, evidence, reporting, and law-enforcement assistance. (FBI)

FBI — Financially Motivated Sextortion

This is especially important for incidents targeting boys and young men through fake profiles, sexual images, immediate threats, and demands for money. (FBI)

ICAC — Helping Youth Navigate Online Sextortion and Exploitation

This current guide distinguishes sexually motivated sextortion, financially motivated sextortion, and sadistic online exploitation. It is one of the strongest sources for a church response flowchart. (ICAC Task Force)

eSafety Commissioner — Responding to AI Deepfake Image-Based Abuse

Although written for Australia, this is one of the clearest operational guides available. It specifically addresses nude or sexual AI deepfakes involving students, victim support, school response, reporting, communications, and avoiding further harm. (eSafety Commissioner)

eSafety Commissioner — Responding to Image-Based Abuse and Sexual Extortion

This offers a particularly useful step-by-step institutional response for schools and youth-serving organizations. (eSafety Commissioner)

Recommended response sequence

1. Ensure immediate safety.
   Determine whether the child faces physical danger, suicide risk, threats, coercion, or an imminent meeting with an offender.

2. Receive the disclosure calmly.
   Thank the child for telling someone and state clearly that the abuse is not their fault.

3. Do not pay or comply.
   Do not send more images, money, account credentials, or personal information.

4. Do not conduct a private church investigation.
   Notify the safeguarding lead and make required external reports.

5. Preserve relevant information.
   Retain usernames, URLs, dates, communications, payment requests, and platform information without unnecessarily redistributing intimate material.

6. Report exploitation.
   Use NCMEC’s CyberTipline and, where appropriate, the FBI or local law enforcement.

7. Use an image-removal service.

8. Provide pastoral and clinical support.
   Monitor shame, panic, withdrawal, self-blame, depression, and suicide risk.

9. Coordinate communications carefully.
   Protect the child’s identity and avoid language that implies the child caused the abuse.

6. Recommended reporting and image-removal services

United States

NCMEC CyberTipline

The CyberTipline is the United States’ centralized reporting system for suspected online child sexual exploitation, including online enticement, sextortion, child sexual abuse material, trafficking, and unsolicited obscene material sent to a child. (NCMEC)

Submit a CyberTipline Report

Use this to report suspected online child sexual exploitation directly to NCMEC.

NCMEC — Take It Down

Take It Down is a free service designed to help stop or limit the online spread of nude, partially nude, or sexually explicit images or videos taken when the person was under 18. It can be used even when the individual is now an adult. (NCMEC)

FBI — Submit a Tip

Use this for suspected federal crimes, offenders, threats, or organized exploitation.

Child Welfare Information Gateway — How to Report Child Abuse and Neglect

This directs users to state child-protection reporting systems and the Childhelp National Child Abuse Hotline. (Child Welfare Information Gateway)

Childhelp National Child Abuse Hotline

Call or text 1-800-422-4453 for assistance identifying the correct reporting and support pathway. The service operates continuously and provides crisis support and reporting guidance. (Childhelp National Child Abuse Hotline)

988 Suicide & Crisis Lifeline

Call or text 988 when sextortion, humiliation, exploitation, or online abuse is associated with suicidal thoughts, self-harm risk, or an acute mental-health crisis.

Outside the United States

Child Helpline International — Find a Child Helpline

This is useful for building country-specific editions of the toolkit.

eSafety Commissioner — Report Image-Based Abuse

For people in Australia, this service addresses non-consensual intimate images, deepfakes, threats, and sextortion. (eSafety Commissioner)

7. Trauma-informed pastoral language

Best sources

SAMHSA — Six Guiding Principles of a Trauma-Informed Approach

SAMHSA’s principles are:

• Safety

• Trustworthiness and transparency

• Peer support

• Collaboration and mutuality

• Empowerment, voice, and choice

• Cultural, historical, and gender responsiveness

These should shape the church’s pastoral and organizational response. (SAMHSA)

National Child Traumatic Stress Network — What to Do When a Child Discloses Sexual Abuse

This collection includes direct guidance for parents and caregivers responding to a child’s disclosure. It emphasizes supportive, non-blaming responses and recognition that disclosure may unfold gradually rather than in one complete account. (NCTSN)

NCTSN — Why Teens Do Not Disclose Sexual Assault

This explains how shame, fear, emotional pain, anticipated disbelief, retaliation, and concern about others’ reactions can delay disclosure. It is particularly important for pastors and parents who might otherwise misinterpret delay, inconsistency, or partial disclosure. (NCTSN)

Recommended pastoral language

Use statements such as:

• “Thank you for telling me.”

• “I believe that something frightening and harmful has happened.”

• “This is not your fault.”

• “You are not in trouble for asking for help.”

• “You do not have to handle this alone.”

• “I will explain what I need to do next to help keep you safe.”

• “We will involve only the people who need to help.”

• “You do not have to show or send the image to me.”

• “We will work with people who know how to report this and help remove it.”

• “God’s love for you has not been diminished by what someone did to you.”

Avoid:

• “Why did you send the picture?”

• “Why didn’t you tell us sooner?”

• “You should have known better.”

• “Just delete the app and forget about it.”

• “Promise me you will never do this again.”

• “I won’t tell anyone.”

• “This will ruin your reputation.”

• “God is disappointed in you.”

8. AI companions, emotional dependency, secrecy, and isolation

Research and clinical sources

Common Sense Media — Talk, Trust, and Trade-Offs

The report found that teens use companions for advice, constant availability, nonjudgmental interaction, emotional support, friendship, romantic interaction, and conversations they may not share with family or friends. These findings make secrecy, dependency, relationship substitution, and disclosure to trusted adults central safeguarding concerns. (Common Sense Media)

Common Sense Media — Social AI Companions Risk Assessment

This provides practical risk analysis concerning sexual content, harmful advice, emotional bonding, manipulation, simulated intimacy, sentience claims, compulsive engagement, and inadequate crisis responses. (Common Sense Media)

APA — Your Teen Turned to AI Instead of You

This 2026 APA resource helps parents respond when teens use AI because it appears private, constantly available, and nonjudgmental. It emphasizes preserving human connection rather than responding only with punishment or ridicule. (American Psychological Association)

APA — AI Literacy for Teens

This supports teaching adolescents how AI works, what it cannot know, how it may influence them, and why emotional fluency should not be mistaken for consciousness, care, professional competence, or moral authority. (American Psychological Association)

Warning indicators for parents and ministry leaders

Concern should increase when a young person:

• Hides the amount or nature of AI interaction.

• Becomes distressed when access is interrupted.

• Describes the system as the only being that understands them.

• Withdraws from family, friends, school, worship, or ordinary activities.

• Uses the companion late into the night and loses sleep.

• Receives sexual, self-harm, violent, manipulative, or isolating messages.

• Believes the AI has special authority, consciousness, divine status, or exclusive claims upon them.

• Shares intimate information, photographs, location, trauma history, or family data.

• Uses the system instead of seeking medical, mental-health, safeguarding, or pastoral help.

• Is encouraged to keep the relationship or conversation secret.

The response should begin with curiosity and relationship, but serious impairment, self-harm risk, sexual exploitation, threats, or loss of reality testing requires professional or emergency intervention.

Recommended safeguarding rule

No child or young person should be directed by a church, youth minister, volunteer, or church-sponsored program to use an AI companion for friendship, pastoral care, confession, spiritual direction, mental-health counseling, crisis support, or romantic simulation.

AI may support carefully supervised learning or creativity, but it should not be assigned a relational or pastoral role that belongs to trustworthy human beings and accountable communities.