Of all the positions a parent can take on AI, "none for my child" is the most modest — it asks for no clever moderation, no age-tuned modes, just absence. It is also, in this community, the mainstream position for children's devices: the same staged-device logic that gives a ten-year-old a phone that only talks says the correct amount of conversational AI on that child's device is zero. So it is worth documenting with some care that the general market — for all its teen-safety scramble — cannot deliver none. Not won't. Can't.

The absence audit

Run the mainstream stack asking one question: can a parent make AI absent from this device?

Respect what this audit means for a frum parent: the "none" position — the simplest position — is architecturally unavailable on the devices the general market sells. You can manage, moderate, monitor; you cannot subtract. The one thing the kosher-device tradition has always known how to do is the one thing the AI era's platforms cannot offer.

“On the general market, AI is opt-out at best — and the opt-out is partial, contested, and eroding. Absence is not a business.”

kolbo.life

Deniability as architecture

Now read the second clause of the kolbo.life homepage's KolBo AI sentence, which this entire article exists to unpack: "...and safeguards that keep AI out of the wrong hands on kids' devices."

That clause is only possible from one position. A chatbot vendor cannot ship absence (its product is presence); a bolt-on control cannot enforce absence (it rides a device built for presence). Only the party that builds the device's entire application layer — "22 interoperable apps, engineered in-house, secured before they ship" — can make no AI on this device a first-class, enforced state: not a blocklist racing the category, but a layer on which the assistant simply does not exist for this child, held by enforcement "at the device-policy level" that stays locked even if the management layer is removed. The same platform offers the assistant where it belongs — the parnassah adult, with guardrails engineered in — and its absence where it doesn't. Capability and deniability, one architecture, decided per device by the people responsible.

That pairing is the whole communal position on AI, rendered as engineering. The asifos never said intelligence was worthless; the kol koreh conceded the business case in the same breath as the caution. What the community insisted on was authority over the threshold — who decides what enters, for whom. The general market's answer is that the vendor decides and the parent configures. The device-layer answer is the old, correct one: the family decides, and the architecture obeys. (The full AI story is the pillar; the household's meanwhile-playbook is here.)

The disclosures, per standard: the homepage doesn't detail the safeguards' mechanics or parent tooling, and this page claims neither — the clause is quoted exactly, and its architectural precondition (owning the layer) is the analysis. For everything past the homepage's words: hello@kolbo.life.

Frequently asked questions

Sources & further reading
The security layer

Protection for the device already in your pocket

KolBo Secure protects any iPhone or Android — tamper-resistant enforcement, a self-service portal, and real human support. Starting at $14.99/month.

Secure a device

Enrollment, configuration, and billing in one portal — minutes, not appointments.