sponto
AI & consent

AI that asks first.

Some of our Apps use artificial intelligence to help you write, summarise, transcribe, create or understand things. This policy explains exactly how those features work, what data they use, and how you stay in control.

Last updated: 29 October 2026Supplements our Privacy Policy
Off by defaultCloud AI features stay off until you say yes.
Clear disclosureWe tell you what's shared, with whom, and why.
No trainingYour content is never used to train AI models.
One-tap opt-outWithdraw consent any time in Settings.

1. Our principles

  • Consent-first. We never send your personal data or content to a third-party AI service without your explicit, informed permission.
  • On-device first. If a task can be done well on your device — using Apple Intelligence, Apple's on-device foundation models, Core ML, Vision, Natural Language or Speech frameworks — we do it there.
  • Data minimisation. We send only what's necessary for the specific request you make, and strip identifiers where we can.
  • Human control. AI suggests; you decide. You can review, edit or discard any Output.
  • Accountability. We choose providers carefully, bind them by contract, and keep this policy accurate.

2. Types of AI processing

TypeWhere it runsData leaves your device?Consent needed?
On-device AI (Apple Intelligence, Core ML, Vision, Speech)Your iPhone, iPad, Mac or Apple WatchNoNo separate consent; system permissions still apply
Apple Private Cloud ComputeApple's privacy-preserving servers, under Apple's termsProcessed by Apple, not by usGoverned by Apple's settings
Third-party cloud AIServers of an AI provider we've contracted (via our secure backend)Yes — only what's needed for your requestYes — explicit opt-in

Before you first use any feature that relies on third-party cloud AI, the App shows a clear consent screen. It will:

  1. explain what data will be sent (for example, “the text you select”, “the photo you choose”, “your voice recording”);
  2. name who will process it (the AI provider, or the category of provider and where more information can be found);
  3. explain why (the specific feature) and how long it may be retained;
  4. ask you to actively choose “Allow” or “Not now”. Pre-ticked boxes and implied consent are never used.

If you choose “Not now”, the rest of the App keeps working; only the AI feature that needs cloud processing stays unavailable. If we later want to share a new type of data, or use a new provider in a materially different way, we'll ask again. We keep a record of when you gave or withdrew consent so we can honour your choice. This approach is designed to meet the App Store Review Guidelines on disclosing and obtaining permission for sharing personal data with third-party AI, and the consent requirements of the UK GDPR, EU GDPR and similar laws.

Example consent message: “To summarise this note, [App] will send its text to our AI provider, [Provider], over an encrypted connection. Your text is used only to produce the summary, isn't used to train AI models, and is deleted by the provider within 30 days. Allow AI summaries?” [Not now] [Allow]

4. What data is shared

Depending on the feature you use, an App may send:

  • Input content you explicitly provide or select — text, prompts, images, documents or audio recordings;
  • Minimal context needed for a good result — such as your chosen language or style setting;
  • Technical data — a random request identifier and rate-limiting token to prevent abuse. We don't send your name, email, Apple Account information, contacts, location or advertising identifier to AI providers.

Requests are routed through our secure backend wherever practical, so the provider receives the request from Sponto rather than from your device and doesn't see your IP address.

5. AI providers

We work only with established AI providers that offer enterprise or API terms meeting our standards. Depending on the App, these may include providers such as OpenAI, Anthropic, Google (Gemini / Vertex AI), and others offering speech, vision or image-generation models. Each App's consent screen and privacy notice names the provider(s) it uses. We require each provider to:

  • process data only to provide the service to us (acting as our processor/service provider);
  • not use our API data to train or improve their models;
  • protect data with industry-standard security and encryption;
  • delete data within a limited period; and
  • support lawful international transfer mechanisms (such as Standard Contractual Clauses).

Naming a provider describes who processes data on our behalf; it does not imply any endorsement, partnership or brand affiliation.

6. No training on your data

We do not use your Input or Output to train, fine-tune or improve any AI model — ours or anyone else's. If we ever wanted to (for example, to improve a feature using voluntarily submitted feedback), we would ask for separate, explicit opt-in consent, and you could say no without losing access.

7. Retention

  • Sponto: we don't store your AI Input or Output on our servers beyond what's needed to return the result, unless you choose to save it (for example, in a history you can delete).
  • Providers: may retain requests for a limited period (typically no more than 30 days) solely for abuse and safety monitoring, after which they're deleted. Where available, we use zero-data-retention options.
  • On device: Output saved in an App stays on your device or in your iCloud until you delete it.

8. Withdrawing consent

You can withdraw consent at any time by turning off AI features in the App's Settings (usually under Settings › AI Features or Privacy). Once withdrawn, no further data is sent to the AI provider. Withdrawal doesn't affect the lawfulness of processing before you withdrew. You can also email [email protected] to withdraw consent or ask for any stored AI data to be deleted.

9. Safety and content rules

To keep AI features safe, Apps may use automated filters (ours and the provider's) on Input and Output. You must not use AI features to create or request content that:

  • sexualises minors, or is non-consensual intimate imagery;
  • is hateful, harassing, violent extremist or promotes self-harm;
  • deceptively impersonates a real person, or creates misleading “deepfakes” of real people;
  • infringes copyright, trademarks or other rights — including generating content using other companies' brands or logos in a misleading way;
  • facilitates illegal activity, fraud, malware or weapons; or
  • contains other people's personal data without a lawful basis.

We may block requests, limit usage or suspend AI access for misuse. You can report harmful Output via the in-app report option or by emailing us.

10. Limitations of AI

AI can be wrong, biased, out of date or confidently incorrect (“hallucinations”). Output is not professional advice. Always check important information and use your own judgement — especially for health, legal, financial or safety matters. See Terms of Use, Section 6.

11. Age requirements

Third-party cloud AI features are intended for users aged 13 and over (or older where local law or the provider's terms require, such as 18 for certain features). They are not offered in Apps in the Kids Category.

12. Transparency and labelling

We clearly indicate in the App when content is AI-generated. Where required by law — such as the EU AI Act's transparency obligations — AI-generated images, audio or video may include a visible label or machine-readable marker, and chat-style features will make clear you're interacting with an AI, not a human.

13. Questions

For questions about AI features, consent, or to request deletion of AI-related data, contact [email protected] or use the contact form.