Acceptable AI Use Policy

Last updated: August 22, 2021

This Acceptable AI Use Policy describes the rules that apply to your use of artificial intelligence, machine learning, and model-powered functionality made available through websites, applications, APIs, software, and related services operated by Anthony Kung ("we", "us", or "our").

This policy is intended to set clear boundaries around how AI-powered features may be used, what categories of activity are prohibited, and what responsibilities remain with the user. It applies whether AI functionality is accessed through a web interface, an API, an integration, or another supported product surface.

AI systems can be powerful and useful, but they also create heightened risks when used irresponsibly. Those risks include generation of deceptive material, facilitation of harassment or abuse, unsafe automation, privacy violations, and over-reliance on probabilistic output. For that reason, use of any AI-powered feature is conditioned on compliance with this policy.

1. Scope

This policy applies to AI-powered or model-assisted features offered as part of the Services, including:

text generation and rewriting
summarization and classification
search augmentation and content extraction
moderation and trust-and-safety tooling
code generation or software assistance
workflow automation and agent-like functionality
image, audio, or multimodal generation where supported

This policy applies to prompts, instructions, context supplied to a model, generated outputs, downstream use of those outputs, and attempts to manipulate or circumvent system behavior.

2. Intended Use

Our AI features are intended to assist with legitimate research, productivity, drafting, analysis, support, moderation, engineering, and creative work. They are designed to augment human judgment, not replace it.

You remain responsible for evaluating whether any output is appropriate for your use case. AI-generated material may be incomplete, biased, misleading, nonsensical, or factually wrong. Even where output appears fluent or authoritative, it should not be treated as inherently accurate or suitable for high-stakes decisions without independent review.

3. Prohibited Uses

You may not use our AI features to engage in, support, or facilitate:

unlawful activity or evasion of legal obligations
fraud, phishing, impersonation, or deceptive practices
harassment, hate, threats, bullying, or targeted abuse
sexual exploitation, including any content involving minors
promotion or facilitation of violence, terrorism, or self-harm
malware development, credential theft, exploit delivery, or cyber abuse
unauthorized collection, disclosure, or misuse of personal information
infringement of intellectual property, privacy, publicity, or confidentiality rights
disinformation campaigns, spam, or manipulative synthetic content intended to mislead others
efforts to bypass safeguards, model restrictions, moderation controls, authentication, or platform limits

The examples above are illustrative and not exhaustive. We may restrict other uses that create material risk to users, third parties, infrastructure, or the public.

4. High-Risk Decision Making

You may not rely on our AI features as the sole basis for decisions in contexts where errors could result in significant personal, legal, financial, educational, housing, employment, or safety consequences.

Examples of high-risk use include:

medical diagnosis or treatment decisions
legal advice or legal rights assessment
lending, insurance, or investment decisions
employment, admissions, or housing eligibility decisions
child safety assessments
law enforcement, surveillance, or public-sector rights determinations
operation of safety-critical infrastructure or equipment

If you choose to incorporate AI outputs into a higher-risk workflow, you are solely responsible for implementing qualified human review and any other safeguards required by law or sound professional practice.

5. User Responsibility for Inputs and Outputs

You are responsible for the content you submit to AI-powered features and for how you use the resulting outputs. That responsibility includes ensuring that you have the necessary rights to submit content, that your use complies with applicable law, and that outputs are reviewed before being acted upon, published, or delivered to others.

You must not represent AI-generated output as verified fact where verification has not occurred. You must not conceal material use of synthetic content where disclosure is required by law, platform rules, contract, or the expectations of the relevant audience.

6. Data Handling and Model Processing

Use of AI-powered features may require processing of prompts, uploaded content, metadata, and generated outputs by us or by service providers acting on our behalf. Depending on the product, data may be temporarily cached, logged, or reviewed to provide the requested functionality, maintain performance, investigate abuse, or improve reliability.

Our handling of personal information is described further in the Privacy Policy. If you are a business customer with privacy compliance requirements, the Data Processing Agreement may also be relevant where applicable.

7. Safety Systems and Monitoring

We may implement technical, operational, and policy controls intended to detect and limit abuse. Those controls may include rate limits, content filters, trust and safety classifiers, heuristic analysis, manual review, and abuse detection based on account behavior or traffic patterns.

You may not probe, disable, or attempt to circumvent these controls. You may not use adversarial prompting, automated prompt variation, or other tactics to intentionally extract restricted content or undermine model safeguards.

8. Enforcement

If we reasonably believe that your use of AI-powered features violates this policy, the Terms of Service, or applicable law, we may take action with or without prior notice. Depending on the circumstances, enforcement may include:

rejecting or blocking prompts or requests
suppressing or removing generated content
restricting account, workspace, or API functionality
lowering rate limits or suspending access
terminating access to affected Services
preserving records and cooperating with legal process where required

Enforcement decisions may be informed by automated signals, manual review, user reports, or third-party complaints.

9. Changes to This Policy

We may update this policy from time to time to reflect changes in law, technology, product functionality, or trust and safety practices. Updated versions become effective when posted unless a later effective date is specified.

10. Contact

Questions about this policy may be sent to support@anth.dev.