AI Usage Policy

1. What AI Features We Offer

ProperResponse currently offers the following AI-assisted features:

We may add AI features over time. Any new AI feature will be documented in this policy before or at the time of release.

2. The AI Technology We Use

ProperResponse's AI features are powered by Claude, a large language model developed by Anthropic, PBC. We access Claude via Anthropic's API.

Key facts about our AI infrastructure:

• Model: Anthropic Claude (version varies as Anthropic releases improvements; we use recent production-ready versions).
• Data transmission: when you invoke an AI feature, the relevant case context you have entered is sent over an encrypted HTTPS connection to Anthropic's API servers in the United States.
• No training on your data: Anthropic's enterprise API does not use submitted prompts or responses to train or fine-tune AI models. Your case data does not become part of any AI training dataset.
• No persistent memory: each AI request is independent. The AI does not "remember" prior conversations between sessions.
• We do not fine-tune on your data: ProperResponse does not use your case content to fine-tune, retrain, or customize any AI model.

3. What AI Can and Cannot Do

4. Your Responsibilities When Using AI Features

By using any AI feature on ProperResponse, you agree to the following:

1. Independent verification: you will independently verify every legal citation, statute reference, case name, rule number, and legal argument generated by AI before relying on it for any purpose, especially in a legal proceeding.
2. Professional review: where possible, you will have AI-generated legal content reviewed by a licensed attorney admitted in the applicable jurisdiction before submitting it to a court or using it in any legal proceeding.
3. Your judgment governs: you understand that AI suggestions are a starting point for your own research and thinking, not a final answer. You are responsible for the final content of anything you file, submit, or communicate.
4. Accurate inputs: you will not deliberately provide false facts or misleading case information to the AI in order to generate content you know to be improper.
5. Confidentiality awareness: you understand that the case context you submit to AI features is transmitted to Anthropic's API. Do not submit information subject to third-party confidentiality obligations you are not authorized to disclose.
6. No reliance for deadlines: you will not rely on AI-generated content to calculate, identify, or plan around court deadlines. See Section 10.

5. How Your Data Is Used in AI Features

When you click an AI feature button, ProperResponse constructs a prompt using:

• The case facts and context you have entered (party names, case type, court, relevant notes or details you have added)
• Your explicit selections or inputs on that specific AI request

What we do not send:

• Documents you have uploaded (file contents are not automatically included in AI prompts unless a future feature explicitly offers document analysis with your opt-in)
• Financial data, Social Security Numbers, or other sensitive identifiers stored in your account
• Other tenants' data
• Your password, 2FA secrets, or any authentication credentials

AI outputs (the responses you receive) are stored as part of your case record so you can reference them later. They are retained subject to our standard data retention policy (see Privacy Policy, Section 8).

For full details on how Anthropic handles data submitted via the API, see anthropic.com/privacy.

6. AI Output Accuracy and Hallucinations

Specific risks to be aware of:

• Fabricated citations: AI may generate case names, docket numbers, and quotations for cases that do not exist or that say something different from what the AI reports. Every citation must be verified in an authoritative legal database (Westlaw, LexisNexis, Google Scholar, CourtListener, or the official court website).
• Outdated law: AI training data has a knowledge cutoff. Statutes may have been amended, cases may have been overruled, and procedural rules may have changed since the AI was trained.
• Jurisdiction errors: legal rules vary dramatically by state, circuit, and court. AI may conflate rules from different jurisdictions or apply federal law where state law applies.
• Overconfidence: AI typically presents information in confident, authoritative language. This tone does not reflect accuracy. Treat AI output the same way you would treat an unverified internet forum post — a lead to research, not a conclusion.

ProperResponse is not liable for any harm arising from reliance on AI-generated content without independent verification. See Terms of Service, Section 5 for the full AI disclaimer.

7. Prohibited Uses of AI Features

You may not use ProperResponse AI features to:

• Generate fraudulent content — fabricated evidence, forged documents, false declarations, or any content intended to deceive a court, opposing party, or any person.
• Harass, threaten, or harm — generate content designed to intimidate, harass, stalk, or harm any individual.
• Violate any law — produce content that constitutes criminal conduct, violates court orders, or assists in fraud.
• Circumvent professional responsibility rules — if you are a licensed attorney, use AI to produce content that would violate the Model Rules of Professional Conduct or equivalent rules in your jurisdiction.
• Train competing AI systems — scrape, extract, or use AI outputs to train, fine-tune, or evaluate any machine learning model or competing product.
• Misrepresent AI authorship — present AI-generated content to a court as entirely your own independent legal work where court rules require disclosure of AI assistance. Check your jurisdiction's rules on AI disclosure obligations.
• Submit to third-party AI services — copy case context or AI outputs from ProperResponse and submit them to third-party AI tools in a way that violates applicable confidentiality obligations.
• Automated bulk requests — use scripts, bots, or automated tools to generate large volumes of AI requests outside normal human use of the application.

Violation of these prohibitions may result in immediate suspension of your account without refund. We reserve the right to report suspected fraudulent use to relevant authorities and courts.

8. AI and the Unauthorized Practice of Law

In many jurisdictions, providing legal advice without a law license constitutes the unauthorized practice of law (UPL), which may be a criminal offense or subject to civil sanction.

ProperResponse does not provide legal advice. Our AI features do not provide legal advice. AI-generated content is a software output — a technology tool — not the professional judgment of a licensed attorney. Using ProperResponse or its AI features does not constitute receiving legal advice, and no attorney-client relationship is formed between you and ProperResponse or any other entity by your use of these features.

If you are a pro se (self-represented) litigant:

• AI-generated legal theories and strategies are starting points for your own research, not conclusions you can rely on without verification.
• Courts may hold pro se litigants to the same procedural standards as attorneys. AI assistance does not excuse procedural errors.
• Some courts are beginning to require disclosure when AI was used in drafting filings. You are responsible for knowing and complying with your court's specific rules on AI disclosure.
• We strongly encourage you to consult a licensed attorney, even for a limited-scope review, before filing any AI-assisted document with a court.

9. AI Features for Legal Professionals

If you are a licensed attorney or other legal professional using ProperResponse:

• Professional responsibility: you are responsible for supervising AI-generated work product as you would supervise a junior associate or paralegal. Model Rule 5.3 (Responsibilities Regarding Nonlawyer Assistance) and equivalent state rules apply.
• Competence: under Model Rule 1.1, competent representation requires understanding the benefits and risks of relevant technology, including AI. Using AI without understanding its limitations may constitute a competence violation.
• Candor: under Model Rule 3.3, you may not make false statements of fact or law to a tribunal, including submitting AI-generated citations you have not verified.
• Client confidentiality: submitting client confidential information to AI features constitutes a disclosure outside the attorney-client relationship. Confirm that such disclosure is permitted under your jurisdiction's professional responsibility rules and your client's informed consent, if applicable.
• AI disclosure rules: some courts and jurisdictions now require disclosure of AI use in drafting. You are responsible for monitoring and complying with applicable disclosure requirements.

10. No Reliance for Court Deadlines

This limitation also applies to the deadline tracking feature in ProperResponse generally. Deadline entries in the system are exactly what you enter — they are not calculated or validated by any AI or automated system. You are solely responsible for the accuracy of deadlines you record.

11. Feedback and Improvements

We want our AI features to be useful, accurate, and responsible. If you encounter an AI output that is factually wrong, potentially harmful, or otherwise concerning, please report it to support@properresponse.com with the case context (removing any truly confidential information) and a description of the issue.

We review feedback to identify patterns of error and improve our prompting approach. We do not use individual feedback to fine-tune the underlying AI model (that is Anthropic's domain), but we use it to improve how we construct prompts and present outputs.

12. Changes to AI Features and This Policy

AI technology is evolving rapidly. We may:

• Add new AI-powered features to the platform;
• Change the underlying AI model or provider;
• Modify the scope of case data sent to AI services;
• Retire existing AI features.

When we make material changes to AI features that affect data handling, we will update this policy and notify account administrators via email at least 14 days before the change takes effect. For non-material changes (e.g., improvements to prompt quality), we will update this policy without advance notice but will update the "Last Updated" date.

13. Contact

Questions about AI features or this policy? Contact us at:

• Email: support@properresponse.com
• Privacy / data questions: privacy@properresponse.com