AI Usage Policy

This AI Usage Policy describes how artificial intelligence is applied within the Narrative AI platform ("Platform"), the limitations of AI-generated content, and the responsibilities of officers and agencies when using AI-assisted documentation tools. This policy is part of your agreement with FirstLine Intelligence Systems, LLC.

1. How AI Is Used in the Platform

Narrative AI uses large language models (currently powered by OpenAI's GPT-4o series) to assist law enforcement officers in drafting incident report narratives based on facts, details, and observations provided by the officer. The AI also powers the Internal Affairs Risk Engine (IARE), which analyzes draft reports for documentation gaps, vague use-of-force language, missing legal justifications, and timeline inconsistencies.

AI is applied in the following features: (a) report narrative generation from officer-provided inputs; (b) narrative enhancement, proofreading, and refinement; (c) pre-submission risk scoring and flagging; (d) auto-fix suggestions for identified documentation issues; (e) case analysis and summarization.

2. Limitations of AI-Generated Content

AI-generated content is a starting point, not a finished product. The AI does not have independent knowledge of the incident, the law, your agency's specific policies, or local legal standards. AI outputs:

• May contain inaccuracies, hallucinations, or fabricated details if the officer's input is incomplete or ambiguous;
• Are not legal advice and should not be treated as such;
• May not reflect current case law, jurisdiction-specific requirements, or your agency's policies;
• Cannot independently verify facts, witness statements, or physical evidence;
• Are not admissible as AI-authored documents — all submitted reports must be authored and certified by a human officer.

3. Officer Responsibility

Every officer using the Platform is solely responsible for: (a) reviewing all AI-generated content before saving, submitting, or sharing any report; (b) correcting any inaccuracies, omissions, or hallucinations in AI-generated drafts; (c) ensuring the final report accurately reflects the officer's firsthand observations and knowledge; (d) complying with agency policy, department SOPs, and applicable law regardless of AI suggestions; and (e) certifying the accuracy of any report submitted under their name or badge number.

No report should be submitted without officer review. The Platform explicitly warns users that AI assists but does not replace officer judgment.

4. IARE — Internal Affairs Risk Engine

The IARE feature is an analytical tool designed to help officers identify potential documentation weaknesses before submission. IARE outputs are advisory only. A low risk score does not guarantee legal sufficiency, and a high risk score does not mean a report is improper. IARE analysis should be one input among many in the officer's report-review process.

5. AI Training and Data Use

We do not use your incident data, report content, or personally identifiable information to train or fine-tune AI models without your agency's explicit written consent. Inputs sent to OpenAI's API are governed by OpenAI's API data usage policies, which prohibit OpenAI from using API inputs to train its models. A current copy of OpenAI's usage policy is available at openai.com/policies.

6. Bias and Fairness

AI language models may reflect biases present in their training data. We encourage officers to critically evaluate AI suggestions for bias, particularly in descriptions of subjects, motivations, or behavior. If you observe consistently biased outputs, please report them to info@narrative-ai.org so we can investigate and address the issue.

7. Audit Logging

The Platform maintains audit logs of AI-assisted actions, including when AI generation was initiated, when IARE was run, and when auto-fix was applied. These logs may be used by agency administrators, internal affairs, or legal counsel to assess how the Platform was used in connection with any report.

8. Updates

As AI technology and legal standards evolve, this policy will be updated. We will notify users of material changes. Continued use of the Platform constitutes acceptance of the current AI Usage Policy.

9. Contact

Questions about AI usage in the Platform? Contact info@narrative-ai.org.