Germany's main Protestant and Catholic church bodies have joined a children's media association in launching an initiative that calls for binding standards on artificial intelligence products used by children, including a code of conduct for developers and new legal obligations for providers.
The initiative, "KI kindgerecht!" ("Child-Appropriate AI!"), was announced Oct. 14 by the Evangelical Church in Germany, or EKD, the umbrella body of the country's regional Protestant churches. Its partners include the EKD, the Church and Society Department of the German Bishops' Conference secretariat, the Protestant media agency GEP and Seitenstark, a network of German children's websites.
The organizers said they are responding to 56 recommendations issued in June by a federal expert commission on protecting children and young people online. In the EKD's announcement, they argued that shielding children from harm is insufficient on its own and that children also need access to good digital services, adult guidance and the skills to use AI themselves. They rejected a blanket ban on AI for minors as unworkable.
"AI is already embedded in many digital applications," said Bishop Georg Bätzing, chairman of the bishops' conference's Commission for Media. He said Pope Leo XIV's encyclical Magnifica humanitas gives the church "a clear mandate" to defend human dignity in the age of AI, "especially where children are easily exposed to dangers."
Bätzing said responsibility "must not be shifted onto parents alone," and named providers, policymakers, educational institutions, civil society and families as sharing it. "Technology must serve people, not purely economic interests," he said.
A code of conduct for interactive AI
The initiative sets out three measures. The first is a voluntary code of conduct for interactive AI aimed at children, which the churches and Seitenstark plan to draft with researchers, educators and interested companies.
According to the EKD, the code will rest on five principles: safety by design, respect for children's rights, age-appropriate spaces for participation, data protection, and human oversight of AI systems.
Building on that code, the partners plan to develop practical quality standards. These are meant to serve as a reference for developers, providers and educational institutions that design or use AI applications for children.
The third measure is a call for legislation obliging providers to meet child-appropriate standards. The partners also want sustained public funding for media education so that children, parents and education professionals can learn to use AI competently.
Participation is open to schools, media providers, researchers, parents' associations, businesses and individuals, the announcement said. It closed with a question: "Will we shape AI for children — or will we leave it to AI to shape our children?"
'The last to be asked'
Darja Martens, a board member of Seitenstark, said young people have largely been left out of decisions about technology they already use.
"Children today are the first to grow up with AI. But often the last to be asked about how it is designed," she said.
Children need protected spaces where they can test, understand and question AI at a level suited to their age, Martens said. Seitenstark will contribute its experience in children's media and from its quality seal for digital children's media, which the group says it has awarded since 2022.
Dorothee Wüst, church president of the Evangelical Church of the Palatinate and chair of GEP's supervisory board, said children "must neither be left alone nor excluded in the digital space." The partners want to help ensure that children can use AI "safely and in a self-determined way," she said.
New risks, yet a ban is not the answer
On its website, the initiative lists the dangers it sees in digital media. They include graphic images of war, targeted disinformation, deepfakes and AI-generated nude images, as well as hate speech, cyberbullying, online grooming and sexualized violence.
Generative AI that interacts with users brings new risks to privacy and data protection, the site says. It also warns that known threats to children's physical and mental health are growing more severe, and that existing approaches to protecting minors in the media are reaching their limits.
The partners describe public debate on children and digital media, including current arguments over social media, as often driven by fear. They point to calls for a blanket ban for minors that followed the commission's recommendations and describe that option as "the seemingly simplest path" and the worst one.
A ban, they argue, would keep children from learning confident, critical and creative media use in a protected setting. Those skills develop only through supervised practice, age-appropriate explanation, clear rules and good examples, according to the site.
The partners also contend that a ban could not be enforced in practice. AI is already built into search engines, learning apps, voice assistants, gaming platforms and social networks, they say.
Parents can act as gatekeepers depending on a child's age but cannot control the entire digital world on their own, the initiative says. It places structural responsibility on platforms and services that use AI, saying they can decide whether to take children seriously.
Children's rights as framework
The initiative grounds its approach in the United Nations Convention on the Rights of the Child, citing children's rights to protection, participation and empowerment and the principle of the best interests of the child. Those rights should be pursued together rather than set against one another, the partners say.
They plan to begin with their own work, which includes children's media and church-run educational institutions ranging from daycare centers to schools.
In drafting the code's criteria, the group is examining chatbots and agentic AI and drawing on projects already being tested. It names Auxie, a chatbot for children developed for GEP's children's website religionen-entdecken.de, as one example of good practice.
The partners list three goals: broader public awareness and skills-building, quality criteria to help children, parents and teachers develop media literacy, and secure funding for high-quality, noncommercial AI services for children. They say the five principles will be made more specific over the coming months with input from those who join.
Ingrid Stapf, a researcher in media ethics and children's rights at the University of Tübingen's International Center for Ethics in the Sciences and Humanities, is also listed as a partner.
The Independent Expert Commission on Child and Youth Protection in the Digital World was appointed in September 2025 by Karin Prien, Germany's federal minister for education, family affairs, senior citizens, women and youth. It presented its 56 recommendations on June 24.
Seitenstark has worked as a registered association since 2007 to promote children's rights online. Organizations and individuals can join the initiative through its website, ki-kindgerecht.de, the EKD said.
