alongkindHuman Alongside AI

Choice · permission · evidence

Our commitment

AI supports. The human decides. Alongkind’s principles guide development; the evidence must show what each feature actually enforces.

Public principles and implementation limits

Human choice comes first

Our design commitments are explicit permission, bounded authority, visible actions, Stop and Revoke, and human control over continuity. These are requirements to implement and test, not a claim that every control is already complete.

What the current model offers

The app lets visitors explore companion designs, choose or create a companion and use available conversation features. Life-stage categories organise the catalogue; they are not age verification or an enforced parent PIN lock. Young Explorer remains a preview and is not approved for unsupervised child use.

Evidence before guarantees

Immutable receipts, independently verified enforcement and complete Stop/Revoke coverage require implementation evidence and failure tests. The Control Evidence Matrix defines those acceptance requirements. Publishing a rule or specification is not proof that the working model passes it.

Feelings are yours to describe

Wellbeing designs use optional reminders and user-initiated check-ins. They exclude automatic emotional judgement and automatic changes based on inferred stress. They do not diagnose or replace qualified human help.

Private memories require care

Deep Drive is a separate private beta for approved accounts. Before storing a photo, consider the privacy and wishes of every person pictured, especially children. Owning a photograph does not by itself settle their privacy rights. Wider access requires a documented assessment of data use and safeguards.

Shared principles, separate projects

The Law Bible is the governing reference for the standalone Companion Forge / Game concept. Alongkind publishes it as a reference. Shared values do not create a technical integration, common data store or shared permissions with that project or TriadNetwork.

Ask, inspect or challenge

Questions about a feature, a claim or your information can be sent to bendyson@alongkind.com. Our aim is to make each feature’s status, limits and evidence understandable before asking people to trust it.