Acceptable Use Policy

Overview

This Acceptable Use Policy (AUP) sets out the prohibited uses of the Crucible platform. By using our service, you agree to comply with this policy. Violation of this policy may result in the suspension or termination of your account.

Prohibited Actions

You agree not to misuse the Crucible service. Specifically, you shall not:

• Reverse Engineering: Attempt to reverse engineer, decompile, disassemble, or derive the source code or underlying models of the service.
• Model Extraction: Use the service to extract parameters, weights, or training data from our AI models or those of our providers.
• Competitive Training: Use outputs from the service to train, fine-tune, or improve any other artificial intelligence models or competing services.
• Unauthorized Resale: Resell, sublicense, or redistribute access to the service without our express written consent.
• Automated Abuse: Use bots, scrapers, or other automated means to access the service in a way that exceeds reasonable usage limits or disrupts the service.

AI-Specific Prohibited Activities

Given the AI-powered nature of Crucible, the following activities are strictly prohibited:

• Prompt Injection: Attempting to manipulate AI outputs through prompt injection, jailbreaking, adversarial prompts, or any technique designed to override, bypass, or circumvent the system's intended behavior or safety mechanisms.
• System Manipulation: Attempting to override, bypass, or manipulate the debate system's safety mechanisms, guardrails, content moderation, or quality gates.
• Output Gaming: Attempting to manipulate the debate process to produce biased, inaccurate, or predetermined outputs, or to circumvent the structured deliberation process.
• Adversarial Inputs: Submitting inputs specifically designed to cause the AI to produce harmful, illegal, misleading, or unintended outputs.
• Data Poisoning: Attempting to corrupt or manipulate calibration metrics, Knight ratings, or any system learning mechanisms through fraudulent feedback or coordinated manipulation.
• Encoding Attacks: Attempting to use encoding schemes (base64, hex, unicode, etc.) or special characters to bypass content filters or inject malicious instructions.

Illegal & Harmful Content

You may not use Crucible to generate, upload, or share content that:

• Is illegal, fraudulent, defamatory, libelous, threatening, or harassing.
• Infringes upon the intellectual property, privacy, or publicity rights of others.
• Promotes hate speech, violence, discrimination, or illegal acts.
• Contains malware, viruses, or other malicious code.
• Is sexually explicit or pornographic.
• Is designed to generate AI outputs that could cause physical, financial, legal, or reputational harm to any person or entity.
• Is intended to produce discriminatory, biased, or unfair recommendations that could disadvantage individuals or groups.
• Attempts to generate advice for illegal activities, regulatory violations, or circumvention of legal obligations.
• Is designed to manipulate or deceive others, including generating misinformation, disinformation, or fraudulent content.
• Involves personal information of third parties without their consent, or attempts to generate outputs that reveal private information.

Restrictions on High-Stakes Decisions

Crucible is designed to support decision-making, not replace professional judgment. The following restrictions apply to high-stakes use cases:

• Medical/Health Decisions: You may not use the service to generate medical diagnoses, treatment recommendations, or health advice that should be provided by licensed medical professionals. AI outputs related to health matters must not be used as a substitute for professional medical advice.
• Legal Advice: You may not use the service to generate legal advice, contract interpretation, litigation strategy, or regulatory guidance that should be provided by licensed attorneys. AI outputs do not constitute legal advice.
• Financial/Investment Decisions: You may not use the service to generate investment advice, securities recommendations, tax advice, or financial planning that should be provided by licensed financial advisors. AI outputs do not constitute financial advice.
• Employment Decisions: You may not use the service to make employment, termination, promotion, or disciplinary decisions without appropriate human review and compliance with applicable employment laws and anti-discrimination regulations.
• Regulatory Compliance: You may not rely solely on AI outputs for regulatory compliance, audit responses, or legal filings without appropriate review by qualified professionals.
• Life-Safety Decisions: You may not use the service for decisions that could directly impact life safety, critical infrastructure, emergency response, or public safety without appropriate human oversight and professional review.

By using the service, you acknowledge that AI outputs are not a substitute for professional judgment and that you are solely responsible for verifying all outputs before making critical decisions. See Section 2 of our Terms of Service for important disclaimers regarding AI-generated outputs.

Security & Integrity

You must not:

• Attempt to bypass or circumvent any security measures, access controls, content moderation, safety filters, or quality gates.
• Interfere with or disrupt the integrity or performance of the service, including the debate system, AI inference, or output generation.
• Perform penetration testing or vulnerability scanning without our prior written permission.
• Share your account credentials with unauthorized third parties.
• Attempt to inject malicious code, scripts, or instructions into prompts, uploaded documents, or any input fields.
• Attempt to exploit vulnerabilities in the debate system, prompt templates, Knight configurations, or output generation mechanisms.
• Attempt to access, view, or manipulate other users' sessions, data, outputs, or account information.
• Attempt to intercept, monitor, or capture data transmitted between users and our service.

Rate Limiting & Abuse Prevention

To ensure fair and reliable service for all users, we implement rate limiting on certain operations. Rate limits are applied per user account (for authenticated users) or per IP address (for guest users) to prevent abuse and ensure system stability.

• You agree not to exceed reasonable usage limits or engage in patterns that suggest automated abuse, scraping, or systematic exploitation of the service.
• We reserve the right to adjust rate limits, implement additional usage caps, or apply access restrictions to prevent abuse and ensure fair service availability for all users.
• Excessive, suspicious, or anomalous usage patterns may result in temporary or permanent access restrictions, with or without notice.
• You may not use the service in a manner that degrades performance or availability for other users.
• Coordinated or bulk activities designed to manipulate ratings, reviews, Knight calibration, or other system metrics are strictly prohibited.
• Creating multiple accounts to circumvent usage limits, access restrictions, or enforcement actions is prohibited.

If you believe your rate limits are too restrictive for legitimate use, or if you have questions about rate limiting, please contact us at hello@roundtablelabs.ai. We may adjust limits for legitimate business use cases on a case-by-case basis.

Prohibited Use of AI Outputs

You are responsible for how you use AI-generated outputs from Crucible. The following uses of AI outputs are prohibited:

• Using AI outputs in a manner that violates applicable laws, regulations, or third-party rights.
• Presenting AI outputs as your own original work, professional advice, or authoritative statements without appropriate disclaimers indicating they were AI-generated.
• Using AI outputs to make decisions that require professional licensure (medical, legal, financial, etc.) without appropriate human review by qualified professionals.
• Using AI outputs to discriminate against individuals or groups in violation of applicable anti-discrimination laws.
• Using AI outputs in a manner that could cause physical, financial, or reputational harm to yourself or others.
• Republishing or distributing AI outputs in a manner that misrepresents their source, accuracy, or reliability.
• Using AI outputs to create or spread misinformation, disinformation, or fraudulent content.
• Using AI outputs to harass, stalk, threaten, or intimidate any person.

You acknowledge that you bear sole responsibility for your use of AI outputs and any decisions or actions taken based on them.

Enforcement

We reserve the right to investigate any violation of this policy. If we determine that you have violated this AUP, we may take action including:

• Removing offending content.
• Suspending or restricting your access to the service.
• Terminating your account without refund.
• Reporting illegal activities to law enforcement authorities.
• Pursuing legal remedies for violations that cause harm to Roundtable Labs, other users, or third parties.

Enforcement actions may be taken without prior notice in cases of serious violations, including but not limited to illegal activity, security threats, or content that poses immediate harm.

Reporting Violations

If you become aware of any violation of this policy, please report it to us immediately at abuse@roundtablelabs.ai.