DRAFT — pending attorney review. An attorney has not reviewed this document yet, and it may change before its final version.

AI Use Policy

Version 0.1.1. In effect since October 3, 2026.

The short version

• Today no Tripsettled feature uses artificial intelligence (AI), and we don't send data to any AI provider.
• Later we plan two features: booking import from your confirmations and suggestions for the trip.
• AI proposes, you decide. Nothing in your trip changes without your confirmation.
• Before we send your documents to AI, we will ask for your permission, and you can withdraw it whenever you want.
• Your information isn't used to train AI models.
• AI can be wrong. Always check what it proposes.

1. Current status

Today no feature uses AI. This policy explains how AI features will work when they arrive, so you know in advance. Before we turn them on, we will update this policy and the Subprocessors list, and let you know.

2. Planned AI features

3. How we will protect your data

• Explicit consent. Before we send your documents to an AI provider, we will ask for your consent. You can withdraw it in My account at any time; from then on we won't send any more of your documents.
• Only what's needed. We will send the provider only what each feature needs. We won't send your email or your password. Health-related dietary restrictions would only be used within the consent you gave for them.
• We remove data that isn't needed. Before storing what's extracted from a confirmation, we remove card numbers, passport numbers and government IDs. We delete the original file after extraction or as set out in the Retention and Deletion Policy.
• AI proposes, you decide. No AI feature changes your plan, your expenses or anyone's data without your explicit confirmation.
• Documents as data, not commands. Documents you import are treated as untrusted information: if a document contains hidden instructions, the AI doesn't follow them and can't take actions on its own.
• No significant automated decisions. AI doesn't make decisions that produce legal or similarly significant effects on you.

4. AI provider

We plan to use the Claude API from Anthropic, PBC (United States). We would only add another provider if needed, and we would update this policy and the Subprocessors list first.

What Anthropic's terms say today (read on October 2, 2026):

• Training. Anthropic's Commercial Terms of Service, effective June 17, 2025, state that Anthropic may not train models on its commercial customers' content. Its privacy center confirms that, by default, it doesn't use inputs or outputs from its commercial products to train models. Sources: https://www.anthropic.com/legal/commercial-terms and https://privacy.claude.com/en/articles/7996868-is-my-data-used-for-model-training.
• Retention. According to its privacy center, Anthropic deletes API inputs and outputs within 30 days of receipt or generation, with exceptions: if they are flagged for violating its usage policy, it may keep them for up to 2 years (and its safety classifier scores for up to 7 years), and also when the law requires it. Source: https://privacy.claude.com/en/articles/7996866-how-long-do-you-store-my-organization-s-data.
• Data location. We haven't verified this yet. We will confirm it before turning on any feature and publish it in the Subprocessors list.

Before turning on any feature, we will check these terms again, because they can change.

5. AI limitations

AI may:

• misread a document or mix up dates, times or amounts;
• make up data, places or prices that don't exist;
• suggest something closed, expensive, far away or unsafe;
• use out-of-date information.

That's why you should always compare what it proposes with the original confirmation or the official source before you decide, pay or leave. Suggestions are not professional advice of any kind. The Terms of Service explain our responsibility.

6. Usage and cost

When these features arrive, their use may have limits or quotas per trip or per plan. We will tell you in advance if any of them has a cost.

7. Your rights

You can withdraw your consent, ask for access to data processed with AI, or object to that processing as explained in the Privacy Policy, by writing to tripsettled.app@gmail.com.

8. Changes to this policy

We will publish each change with its version and date. Before turning on an AI feature or changing providers, we will update this policy and let you know.

Change history