Independent nonprofit
Keep AI Safe
An independent nonprofit that tests and rates the AI products children use.
- Bob’s role
- Co-founder
- Co-Founder & CEO
- Yarin Kirchen
- Chairman of the Board
- I. Jeffrey Pheterson
- Status
- Active · in its second year
- Visit
- keepaisafe.org
Why it matters to him
Children's safety in the age of AI is, in his words, “the very reason I came out of retirement.” He co-founded Keep AI Safe and is one of its principal funders; its judgments are its own.
What it is
Keep AI Safe Foundation, Inc. is an independent nonprofit, incorporated in Florida in 2025. It is led by its co-founder and chief executive, Yarin Kirchen, and governed by its own board. Its tagline is plain: “Make AI safe for the kids growing up with it.” Its mission is to protect children and adolescents from AI-related harm, through awareness, research and systemic change.
The foundation puts the gap this way: “There is no FDA for AI toys, no NHTSA for AI companions, no Consumer Reports for child-facing chatbots. We are building those functions: independently, publicly, and on a nonprofit footing.” One rule underwrites that independence: “We accept no funding from AI product manufacturers.”
How it works
The work runs as a loop with four parts: test, publish, certify, train.
- Test. Through VerifAI, the foundation runs simulated child conversations against real products: the AI toys, chatbots and companion apps children actually use. A test can show where a product fails; it cannot prove a product safe.
- Publish. Results are scored against StandardAI, its safety criteria for child-facing AI, across seven categories of harm, and published as Toy Ratings and an AI Incident Dashboard.
- Certify. Products that meet the bar will be able to pursue CertifAI, the foundation’s planned “Safe for Kids” mark.
- Train. KAIS Academy, free AI-safety training for families, educators and youth-serving professionals, is listed as coming soon, alongside an early-career practicum.
For a parent, the concrete offerings today are the published product reports and the free parent tools at keepaisafe.org.
Where it stands
The foundation is in its second year. Its product reports and ratings are published on keepaisafe.org. It is listed as an affiliate member of IASEAI and names everyone.AI’s iRaise International Coalition among the organizations it works with.
In October 2026 Bob proposed a next step to the foundation: that it host a private call among the leading children’s AI-safety organizations in the United States, to see whether they can speak with one voice on a single shared red line: “no AI companion or chatbot in direct contact with children under 13 unless it has been independently tested as safe.” His ground rules for such a call: “We host and serve. We do not lead,” and “No AI company money funds this effort.” It is a proposal. A draft invitation exists, and there is no record yet that any organization has been asked.
Bob’s part
Bob co-founded Keep AI Safe with Yarin Kirchen and is one of its principal funders. Yarin is CEO; I. Jeffrey Pheterson chairs the board. What to test, how to rate it and what to publish are their decisions, not his.
A plain disclosure belongs here. Bob is also a co-founder of Boundlia, a company that builds AI for children’s products. Keep AI Safe’s testing and ratings are independent of his companies, and this site makes no claim that it will certify any product of his. From the first talks about a children’s AI company, in the spring of 2026, he asked that any certification and consumer guidance be kept “completely arms length” from it.