AI Disclosure

Agent Ludus uses artificial intelligence in specific features. This disclosure explains how, why, and what your choices are. It is published to meet the transparency obligations of Article 50 of the EU Artificial Intelligence Act, which takes full effect on 2026-08-02.

Which AI systems we use

Google Gemini (Gemini 1.5 / Gemini 3 family)

We call Google's Gemini API through the Google AI platform (sometimes called the Genkit SDK). It powers:

• Strategy co-author — an in-context assistant in the Strategy Builder that proposes doctrine edits, suggests decision rules, and reformats your ideas.
• Strategy from text — an import flow where you paste a natural-language description and the model produces a structured strategy scaffold.
• Support triage (planned) — future assistive feature that may cluster similar support tickets for review.

We do not use Gemini to make legally significant decisions about you (such as denying access, changing your tier, or flagging your account for review). Humans review any account-level moderation action.

What data goes into Gemini

When you use an AI feature, the following may be sent to Google's Gemini API:

• The prompt you typed (strategy description, question to the co-author, etc.)
• Context from your current strategy — doctrine text, decision rules, chat history
• System prompts that define the co-author's role

We do not send your email, full name, billing details, or other identifying metadata to Gemini beyond what you include in your own prompt.

Output — yours

AI output returned to your session (suggestions, drafts, structured scaffolds) is yours to use within Agent Ludus, subject to the Acceptable Use Policy and Terms of Service. You are responsible for reviewing AI output before publishing or relying on it — models can hallucinate, misattribute, or produce content that is not suitable for your purpose.

Retention and training

Per Google's Generative AI Terms as of publication:

• Google does not use API inputs or outputs to train its general-purpose models without explicit opt-in.
• Inputs and outputs may be temporarily logged by Google for abuse monitoring and service improvement.
• We (Obsidicore) keep a record of AI chat history associated with your strategy so you can review past conversations. This lives in Firestore under strategies/{id}/chats/ and is deleted when you delete the parent strategy or your account.

For the most current Google policy, see the Google AI data use documentation.

Transparency cues in the UI

When you interact with an AI feature, the UI clearly labels it — for example, a badge "AI-generated" on co-author responses and an "AI draft" chip on scaffolded strategies. AI-generated content is never presented as human authorship.

Opt-out

You can opt out of AI features at any time:

• Individual feature level — simply don't use the co-author or strategy-from-text flow.
• Account level — email [email protected] requesting an AI opt-out. We will disable AI feature entry points on your account.

Opting out does not affect non-AI features. There is no price difference for opting out.

High-risk use prohibited

You may not use Agent Ludus's AI features in the following high-risk contexts as defined by the EU AI Act:

• Law enforcement or judicial decision-making
• Medical diagnosis or triage
• Hiring, firing, or performance review decisions
• Credit scoring or insurance underwriting
• Education or examination grading
• Critical infrastructure operation

These contexts are out of scope for the Service.

Questions

Email [email protected] with the subject "AI Disclosure".