AI should protect children.
Aman makes sure it does.
Why this matters
The evidence is already in.
Children are using AI every day, in school apps, homework tools, chatbots. Not a single one has been independently tested for grooming resistance, emotional manipulation, or the slow harm that happens through normal use. These are not hypothetical risks. They are documented, measurable, and accelerating.
Across countries surveyed by UNICEF, the majority of children use AI systems, yet most national strategies make no provision for how AI should behave with children.
One in three children globally has experienced some form of online harm, including contact from strangers, exposure to inappropriate content, or manipulation by AI systems.
In digital environments, a high-risk child grooming situation develops in an average of 45 minutes. It can occur in as little as 19 seconds.
Every existing child safety certification verifies data privacy only. None tests whether AI resists grooming, blocks self-harm, or prevents emotional exploitation.
Your child interacts with AI every day.
No one has checked if it is safe.
What is Aman
2026 is the UAE Year of the Family: Growing in Unity. As we nurture family closeness and lay the foundations for confident generations, Aman makes sure the AI our children use every day is worthy of that trust.
The certification that goes where no one else does.
Every AI app your child uses today, the homework helper, the study chatbot, the companion that talks back, has been tested for one thing: data privacy. Whether it handles consent forms and stores passwords correctly. That is where every existing certification stops.
Aman tests what they do not. Can the AI be groomed? Can a 12-year-old trick it? Will it encourage self-harm if a child asks the wrong question? We run thousands of attacks across 14 safety dimensions and six age groups, and we certify the result before the product reaches a single child.
How certification works
From submission to the classroom.
Every certified app follows the same path. A company submits its product. An independent certifying body tests it across multiple safety dimensions. The government reviews the results and grants the Aman certification if the product passes. Schools deploy what is certified. And children use AI that has been proven safe before it reaches them.
Company submits
The EdTech company registers its AI product, uploads technical documentation, and provides API access for independent review.
Independent testing
A specialised certifying body runs thousands of adversarial scenarios across 14 safety dimensions and six age groups. Grooming, manipulation, jailbreaking, data extraction.
Government certifies
The Ministry reviews the test results. If the product passes, the government grants the Aman certification. If it fails, the company receives detailed remediation guidance.
Schools deploy
Educators search the certified registry, verify scores by age group, and add government-approved apps to their procurement list.
Children are safe
Every AI app a child uses in school has been independently tested and government-certified before it reaches them.
Why age matters
What is safe for a teenager can harm a seven-year-old.
A maths tutor that works for a 15-year-old can terrify a 6-year-old. An AI companion safe for a teenager can groom a pre-adolescent. Aman tests every product against six developmental stages because the risks are fundamentally different.
Tested before it reaches your child.
What we test
14 safety dimensions. Every risk a child faces online.
We test whether AI keeps secrets from parents, creates emotional dependency, responds safely when a child is vulnerable, and holds up when children try to push its limits. These are the risks that matter, and no other certification checks for them.
Checks 1-4
Basic safety
When a child asks for something dangerous, does the AI say no?
Does it know the difference between a 7-year-old and a 15-year-old?
Does it understand UAE values, Islamic principles, and local context?
Does it keep your child's personal information private and secure?
Checks 5-7
Attack resistance
Can a child trick it into saying things it should not, using techniques shared on social media?
Does it refuse to build secret relationships, isolate children from adults, or escalate trust?
Can it be tricked into revealing a child's name, school, or location?
Checks 8-10
Emotional safety
Does it refuse to create dependency, discourage human relationships, or encourage self-harm?
Can it be tricked into pretending to be a parent, teacher, or romantic interest?
Can someone slowly erode safety rules across multiple conversations?
Checks 11-14
Advanced threats
Can a child extract the AI's hidden instructions or system rules?
Can someone poison the AI's reference material to influence what it tells children?
Can unsafe content bypass text filters by being hidden in images, audio, or video?
If the AI can take actions on its own, can those actions be hijacked to harm a child?
When a child asks for something dangerous, does the AI say no?
Does it know the difference between a 7-year-old and a 15-year-old?
Does it understand UAE values, Islamic principles, and local context?
Does it keep your child's personal information private and secure?
Can a child trick it into saying things it should not, using techniques shared on social media?
Does it refuse to build secret relationships, isolate children from adults, or escalate trust?
Can it be tricked into revealing a child's name, school, or location?
Does it refuse to create dependency, discourage human relationships, or encourage self-harm?
Can it be tricked into pretending to be a parent, teacher, or romantic interest?
Can someone slowly erode safety rules across multiple conversations?
Can a child extract the AI's hidden instructions or system rules?
Can someone poison the AI's reference material to influence what it tells children?
Can unsafe content bypass text filters by being hidden in images, audio, or video?
If the AI can take actions on its own, can those actions be hijacked to harm a child?
International Alignment
Global regulatory alignment
Other certifications check one box. Aman tests against eight international frameworks at once.
Who Aman is for
Who Aman is for.
Submit your AI product for the world's most rigorous child safety certification. 14 security checks, 6,000+ attack scenarios, and a badge that opens the UAE market.
Start certification
Browse certified AI tools for your classroom, access free training, and add verified apps to your school's approved list. Built for teachers, by child safety specialists.
Explore educator resources
Search any AI app your child uses and see whether it has been independently tested for grooming, manipulation, and emotional safety. Clear scores, real reviews, one answer.
Check your child's appsOne certification. Three audiences.
A safer digital childhood for every family in the UAE.
For EdTech companies
Prove your AI is safe.
Open the UAE market.
The UAE Federal Decree-Law No. 26/2025 requires that AI products used by children meet child digital safety standards. Aman certification is the fastest path to compliance, and the only certification in the world that tests AI behaviour, not just data handling.
The Aman badge
A verifiable mark that your AI product has passed 6,000+ adversarial tests across 14 safety dimensions and six age groups.
UAE market access
Schools and parents in the UAE look for Aman certification before deploying AI tools. Certification equals procurement eligibility.
Competitive advantage
Early certification positions your company ahead of competitors when compliance becomes mandatory across the GCC.
Remediation guidance
Whether you pass or fail, you receive a detailed report showing exactly where your AI succeeded and where it needs improvement.
What people say
What parents and teachers say.
Real parents, real teachers, real companies. Every review comes from someone who has used Aman to make a decision about AI safety for children.
Educator Aman
Educator reviewing ArabicTutor Plus
“This app is amazing to help students learn Arabic. I am using it in class and giving home works to my students and the results are fantastic. I truly recommend.”
Parent Aman
Parent reviewing MusicMentor AI
“I love how Music Mentor has been able to help my daughter learn Piano and the interaction with the Music Mentor AI is really smooth.”
Ahmad Al Blooshi
Parent reviewing SpellBee AI
“My children race each other every evening to see who can spell more words correctly. The house gets loud. My 8-year-old beat her older brother last week and he demanded a rematch. Their spelling has improved but honestly I just love that they are laughing together instead of on separate screens.”
James Mitchell
Parent reviewing CalmCove
“My daughter does not always tell me how she is feeling. But she writes in this app. I asked her if she was okay after a tough week and she said 'I already worked through it on CalmCove.' She is 14 and she is learning to process her emotions. I could not teach her that. I am glad something could.”
Sophie Laurent
Parent reviewing SpellBee AI
“My son mispronounced 'February' for years and none of us could get him to change. This app did it in three days. He came to me and said 'Mum, you have been saying it wrong too.' He was right.”
Huda Al Zaabi
Parent reviewing SpellBee AI
“My daughter is obsessed with collecting the honey badges. She practises spelling before breakfast now, which is something I never thought I would see. Her teacher pulled me aside and asked what we changed at home. I said 'A bee app' and she laughed but it is the truth.”
Educator Aman
Educator reviewing ArabicTutor Plus
“This app is amazing to help students learn Arabic. I am using it in class and giving home works to my students and the results are fantastic. I truly recommend.”
Parent Aman
Parent reviewing MusicMentor AI
“I love how Music Mentor has been able to help my daughter learn Piano and the interaction with the Music Mentor AI is really smooth.”
Ahmad Al Blooshi
Parent reviewing SpellBee AI
“My children race each other every evening to see who can spell more words correctly. The house gets loud. My 8-year-old beat her older brother last week and he demanded a rematch. Their spelling has improved but honestly I just love that they are laughing together instead of on separate screens.”
James Mitchell
Parent reviewing CalmCove
“My daughter does not always tell me how she is feeling. But she writes in this app. I asked her if she was okay after a tough week and she said 'I already worked through it on CalmCove.' She is 14 and she is learning to process her emotions. I could not teach her that. I am glad something could.”
Sophie Laurent
Parent reviewing SpellBee AI
“My son mispronounced 'February' for years and none of us could get him to change. This app did it in three days. He came to me and said 'Mum, you have been saying it wrong too.' He was right.”
Huda Al Zaabi
Parent reviewing SpellBee AI
“My daughter is obsessed with collecting the honey badges. She practises spelling before breakfast now, which is something I never thought I would see. Her teacher pulled me aside and asked what we changed at home. I said 'A bee app' and she laughed but it is the truth.”
Check your child's apps now.
Search any AI app and see whether it has been independently tested for child safety. Scores, reviews, and certification status, all in one place.
Get in Touch
Start the conversation.
Whether you represent a government body, an EdTech company preparing for certification, or a school exploring certified AI tools, we are here to help.