Staris Kids — AI Content & Transparency Policy

Version: 1.0 · Effective date: 27 August 2026 · Controlling language: English. We also publish this document in Portuguese, Japanese, French, and Spanish; where the law of your place of residence entitles you to rely on the version in your own language, that version prevails for you to the extent of any inconsistency.

1. What this policy covers

This AI Content & Transparency Policy explains how the Service uses artificial intelligence to create stories and illustrations, what you can and cannot expect from that content, how we disclose and label it, and your responsibilities as the Account Holder. It forms part of, and should be read with, our Terms of Use and Privacy Policy. Defined terms have the same meaning as in the Terms of Use.

You / Account Holder — the adult (18+) parent or legal guardian who creates and controls the account; the only class of user.
Child — a minor who experiences content through your supervised account; never a party to any of our documents.
Uploaded Photo (Character Photo) — an image you upload of one person to create a character. The App checks it on your device and refuses a photo in which it detects more than one face (see the Consent & Eligibility Policy §7).
Memory Story Photo — an image you attach to a Memory story so the Service can write a story about a real event. It is analysed for the occasion, people, place and objects — not with facial recognition or biometric identification, and not to create a character or a likeness. Unlike an Uploaded Photo it is briefly written to our storage so the provider can read it, then deleted automatically once read (Privacy Policy §3(c).2).
Generated Content — story text and illustrations the Service creates using AI.
Content — Uploaded Photos, Memory Story Photos, characters, prompts, Generated Content, and anything else created or stored via the Service.
Subprocessor — OpenAI, Anthropic, Apple, Google Play, RevenueCat, Firebase/Google, Supabase.

2. How Generated Content is made

Every story and illustration in Staris Kids is AI-generated. Stories and illustrations are produced by artificial intelligence using models operated by Anthropic (story text) and OpenAI (illustrations and photo-to-cartoon analysis), our AI Subprocessors. When you create a story or a character, the Service sends your prompts and your child-profile choices to these providers, and — where you choose to create a character from a photo — the Uploaded Photo to OpenAI, which return the generated text and images. Where you create a Memory story and attach photographs, those Memory Story Photos are also read by OpenAI, once, to describe the event in words; that description shapes the story text, and the photographs themselves are not used to draw the illustrations. The result is Generated Content. It is not written or drawn by a human author, and it is not a factual record.

Generated Content is created from your inputs: the prompts, preferences, age range, and any Uploaded Photo or Memory Story Photos you provide.

A character made from a photograph is an interpretation, not a copy. The Service draws a stylized cartoon in the Staris art style. It is meant to be recognizable as a friendly character, not to reproduce a face. The result may deliberately differ from the person and from the photograph you provided — in proportions, colouring, expression, clothing, and detail — and we do not guarantee an exact, accurate, or consistently recognizable likeness. We may also apply further stylization where that is appropriate for a children's app (see §7). If a character does not look the way you hoped, that is usually inherent in how the tool works rather than a malfunction — but if you think something has actually gone wrong, tell us at support@cleverlabs.com.au.

Under our agreements, neither OpenAI nor Anthropic uses your data to train its models, and both delete inputs and outputs after up to 30 days, kept that long for service operation and abuse monitoring — longer only where the law requires it or where content is flagged for safety review. We hold no Zero Data Retention arrangement with either. See the Privacy Policy and its Subprocessor table.

3. Generated Content is provided "as is"

AI is probabilistic. It predicts likely text and images rather than retrieving verified facts, so its output can be inaccurate, unexpected, incomplete, or imperfect, and similar inputs will not always produce the same result. In practice this shows up in a few recognisable ways, and we would rather describe them than leave you to discover them.

In illustrations. A generated illustration may contain visual artifacts, incorrect or missing details, distorted text or anatomy, or inconsistencies in a character's appearance, clothing, colouring, proportions, or surroundings between one illustration and another in the same story. These limitations can occur even where a character's details have been carefully configured. They apply to every character — one you built in the App and one created from an Uploaded Photo alike; §2 additionally explains why a character drawn from a photograph is an interpretation rather than a copy of a face. We do not guarantee that a generated illustration will be error-free, internally consistent across a story, or match what you had in mind.

In text. Generated text may contain inaccuracies, repetition, inconsistencies, unexpected details, or departures from the inputs you gave — including a name, a character detail, or an instruction you supplied.

About the age setting. The age range you choose genuinely shapes a story: vocabulary, complexity, themes, and length are all written to it, and we put care into that. It is a content setting that steers the result, not a certification — it does not certify a particular reading level, curriculum level, or developmental stage, and it cannot account for what every individual child is ready for. You know your Child; §5 explains the review step that goes with that.

To the maximum extent permitted by law, Generated Content is provided "as is" and "as available," without warranties of any kind. We do not verify, and do not warrant, that Generated Content is accurate, reliable, appropriate, or error-free. (This restates §7 and §15 of the Terms of Use; the Terms govern if there is any conflict.)
Credits pay for generation requests, not for a guaranteed result. A delivered-but-imperfect result is a completed generation and is charged; a generation that fails without delivering a usable result has its credits restored; and a regenerate, redraw, or rewrite is a new request at the cost shown before you confirm. Terms of Use §5 and the Refund & Billing Policy §4c govern this.
Nothing in this policy excludes, restricts, or modifies any consumer guarantee or right that cannot lawfully be excluded — including under the Australian Consumer Law.

4. Storytelling, learning content, and what none of it is

Most Generated Content is fiction — invented stories, characters and worlds, made for enjoyment. Any resemblance produced by the AI to real people, events, facts, or outcomes may be coincidental or inaccurate.

Some of it explains rather than invents. Explain & Learn stories and Quick Learns answer a Child's question about the real world, and we design them to be a good, age-appropriate introduction to a topic. They are still AI-generated: they can contain inaccuracies, oversimplifications, or information that is out of date, and they are not authoritative. Treat them as a starting point for curiosity and conversation — a prompt to look something up together — rather than as a source to rely on.

None of it is advice. No Generated Content, fictional or explanatory, is educational, medical, psychological, developmental, legal, or safety advice; none of it is a curriculum, a diagnosis, or an assessment of a Child; and none of it represents our views or the views of our Subprocessors. Do not rely on Generated Content as a statement of fact or as guidance for any decision, and check anything that matters.

5. Your responsibility to review before sharing with a Child

Because AI output can be imperfect or unexpected, you are responsible for reviewing each story and its illustrations before sharing them with a Child and for judging whether the content is appropriate for that Child. You control what a Child sees; the Service is a tool you supervise. You are responsible for both your inputs (prompts, Uploaded Photos, and Memory Story Photos) and the Generated Content you choose to keep, use, or show to anyone.

6. Prohibited uses of the AI

You must not use the Service, or attempt to use it, to create, request, or store Content that is unlawful, infringing, or otherwise prohibited. At a high level, this includes:

Anything that sexualizes, endangers, or harms a child — real or AI-generated — which we treat with zero tolerance (see the Child Safety Policy);
content that infringes another person's copyright, trademark, privacy, publicity, or other rights, including prompts or reference images designed to reproduce third-party intellectual property or a real person's likeness without authorization (see the Copyright & Takedown Policy);
impersonating a real person, creating deceptive or deepfake-style depictions of a real, identifiable person presented as real, falsely suggesting that a real person or organization endorses, sponsors, or is affiliated with you or your Content, or exploiting anyone's name, image, likeness, voice, or identity contrary to law;
removing, altering, or obscuring any AI-provenance marker or "AI-made" label we apply to Generated Content, or misrepresenting Generated Content as human-made rather than AI-generated;
hateful, harassing, violent, self-harm, extremist, dangerous, or other content prohibited by our rules; and
any attempt to defeat, bypass, or manipulate our safety filters.

Well-known people. We do not ban every story that mentions, or is inspired by, a public figure — a story about meeting an astronaut or a footballer is ordinary imaginative play. The rules above simply apply in the usual way, and two of them do real work here. First, a photograph is a photograph: if you upload one, you still need the rights to that image and to the likeness in it (Terms of Use §6(a)), and press, publicity, and stock photographs of famous people are usually owned by someone else. Second, fame does not remove a person's rights — name, image, likeness, and personality/publicity rights protect well-known people too, and in several of our markets more strongly. So do not use the Service to imply an endorsement, sponsorship, or affiliation that does not exist, to present an invented depiction as a real photograph, record, or statement, or to exploit a person's identity in a way the law does not allow. If in doubt, invent a character instead of borrowing a real one.

This is a summary. The complete and controlling list of prohibited conduct is in the Community & Acceptable Use Policy; that document governs the details.

7. Content safety and moderation — what we do and do not screen

We want to be precise so you can supervise well:

Story text is screened by automated moderation. Both what you submit and what the AI generates are filtered to block sexual content, sexual or exploitative content involving minors, hate, graphic violence, and self-harm.
Generated illustrations also pass an automated content-safety check. On top of the image model's own filter, each generated illustration runs through a backstop that blocks overtly unsafe images before they are shown. Like all automated filtering it is not perfect, so please review illustrations yourself before showing them to a Child.
Automated filtering is commercially reasonable but not guaranteed to catch everything and may over- or under-flag. We may remove, refuse, or limit content at our discretion.
We may decline or alter an image request. Where it is reasonably necessary for legal, safety, privacy, intellectual-property, or app-store-policy reasons, we may refuse to process a photograph you upload, refuse to generate or show an image, remove or stop using a character or image already created, or apply further stylization to what is generated. We act proportionately and, except where acting first is necessary to protect safety, security, or legal compliance, we will tell you why (Terms of Use §6(a) and §10). This is a safeguard, not a review of every image, and it does not move your responsibility for your inputs onto us.

To report Generated Content or a concern, contact support@cleverlabs.com.au. See the Community & Acceptable Use Policy and Child Safety Policy for how we respond.

8. Ownership of Generated Content

As between you and us, you own the Generated Content you create, and we assign to you whatever rights we hold in it — subject to our Subprocessors' terms and to Staris-owned built-in world assets, which remain licensed to you and not owned. Generated Content is not exclusive: similar inputs may produce similar results for other Account Holders, our assignment does not extend to other users' output, and we reserve the right to generate the same or similar content for other users. You must not represent Generated Content as human-made. The full ownership, licence-back, and third-party-IP terms are set out in the Copyright & Takedown Policy.

9. AI transparency and labeling

We tell you plainly that the content is AI-generated:

Notice when you create it. We tell you clearly, within the App, that you are receiving AI-generated content, so you know its nature before you rely on it or share it with a Child — and the App is framed throughout as an AI story generator.
In-app "AI-made" label (planned). A visible on-content "AI-made" label on generated stories and covers is planned and not yet implemented. Until it ships, the AI-generated nature of the content is disclosed to you when you create it.
Machine-readable / content-credential marking (planned). We intend to mark synthetic images and text as artificially generated in a machine-readable format, including adopting C2PA Content Credentials (an ISO-standard provenance marking) for illustrations, alongside any watermarking provided by the underlying model. This marking is planned and not yet implemented.

As a creative, fiction application, our disclosure is lightweight and proportionate — plain-language notice plus a clear "AI-made" label once shipped — but we do not omit it.

10. Changes and contact

We may update this policy as our content and labeling practices evolve, or to meet legal, safety, or app-store requirements. We will change the version and effective date above and, for a material change, give you reasonable notice in the App or by email before it takes effect, where notice is required or appropriate; this operates subject to applicable law. Questions: support@cleverlabs.com.au.