Dozens of Countries Just Pledged That No AI Reaches Children Until It's Tested — but the Pledge Names No Tester and No Enforcer
Back to blog
Family Safety·NOVEMBER 11, 2026

Dozens of Countries Just Pledged That No AI Reaches Children Until It's Tested — but the Pledge Names No Tester and No Enforcer

The Blaick Team 5 min read

On November 10, at the Paris Peace Forum, dozens of countries backed the Children First Declaration, promoted by French President Emmanuel Macron. Its core promise: no AI should be placed in the hands of children without first being "tested, evaluated and therefore authorised."

The number to hold: 2045. That is the year signatories aim to show "tangible results." The pledge is voluntary, and reporting describes no sanctions, no deadlines and no named body to do the testing. Your child's chatbot is not waiting for 2045.


The pledge
Test first
No AI for children until it is tested, evaluated and authorised.
Research funding
€8 million
Pledged by AI developers including OpenAI and Anthropic, plus AI safety groups, to study AI and child development.
Five focus areas
Health to well-being
Health, nutrition, education, protection and psychosocial well-being in the digital age.
What's missing
No enforcer
No named tester, test standard, approval body or penalties in any report we reviewed.

Sources: The National (November 10, 2026); Tech Times (November 11, 2026), citing the Élysée and other reporting. Outlets disagree on the exact number of signatories, and no full list has been published.


A Big Promise With a Small Print Problem

Macron called the declaration a "major step forward," and the money is real: the €8 million is meant to help researchers understand how AI affects children's development. But the Élysée itself calls the commitments voluntary, and its own roadmap is softer than the headline. It says AI that benefits children should, "over time," go through pre-authorisation.

The open questions

Coverage does not say whether general-purpose chatbots, the ones kids actually use, would fall under pre-authorisation. Independent access to the models for testers is described as a goal, not an existing right.

A promise to test AI before children use it is only as strong as the person doing the testing. Right now, nobody has been named.

Where Real Rules Already Exist

Binding, not voluntary

New York's chatbot law for minors takes effect January 1, 2027 and requires age verification. The European Commission president has floated a European Kids Act for AI, with a 2028 target. Those carry penalties; the Paris pledge does not.

The gap for your family

Whatever gets agreed in Paris, the AI in your child's pocket, school laptop and favourite apps is already live. Until independent testing exists, parents are the testing layer.


What You Can Do Now

Test it yourself Spend ten minutes asking your child's chatbot the awkward questions they might ask. See how it answers.
Check the age settings Use teen or child modes where offered, and confirm your child's real age is on the account.
Make AI a family topic Ask what they use it for and what it has said that surprised them. Keep it curious.
Don't wait for 2045 Treat any AI tool as untested for children until you have seen how it behaves for yourself.

Governments are finally saying the right thing: test AI before children use it. Until someone is named to do the testing, the safest tester is a parent who is paying attention.


Be the safety check while the rules catch up

Start a free 14-day Blaick trial — no credit card required.

Blaick helps you see what AI is doing in your child's world, so you don't have to wait for regulators.

Start free trial →
Cancel anytime · Trusted by security-conscious families · Built on enterprise AI security