Usage Policy

Effective: October 5, 2026

The mission of Reflection AI, Inc. (“Reflection”) is to make intelligence open and accessible to all. That mission shapes this policy, which governs your use of our products and services, and any other services that link to it. Please read this policy carefully before using our services.

This policy is intended to set out uses we consider unacceptable and do not allow, plainly enough that you can tell where the lines are before you build and deploy. It is not exhaustive. We may treat other uses as violations where they present risks comparable to those described here.

If a use is not addressed here and you are not sure whether it is consistent with this policy, ask us at aup@reflection.ai.

This policy is in addition to the other terms that govern your access and use of our products and services. Reflection works with many types of organizations, governments, and public-sector customers, with different needs. Where you have a written agreement with us that addresses acceptable use, that agreement supersedes this policy.

Where we release products and services under an open license that does not cross-reference this policy (or similar terms), that open license controls.

Prohibited uses

You agree that you will not directly or indirectly use our products and services to do any of the following, and will not attempt to do so, assist anyone else in doing so, or permit anyone using your access to do so:

Break the law or harm people

• violate any applicable law;
• engage in, promote, glorify, or otherwise materially support violence or terrorism;
• engage in, promote, or otherwise materially support human trafficking, exploitation, or sexual violence;
• engage in, promote, or otherwise materially support sexual solicitation;
• generate, distribute, or threaten to distribute intimate or sexually explicit depictions of a real person, including by altering an existing image or likeness of them;
• harass, abuse, threaten, dehumanize, or bully any person or group;
• promote or incite discrimination or violence against people based on race, ethnicity, national origin, religion, caste, disability, sex, gender identity, sexual orientation, or other characteristics protected under applicable law;
• engage in or promote discrimination based on characteristics protected under applicable law, or engage in other unlawful conduct, in the provision of employment, credit, housing, other economic benefits, or other essential goods and services;
• practice a licensed profession without the license or authorization to do so; or
• infringe, misappropriate, or otherwise violate the rights of a third party.

Endanger minors

In this policy, “minor” means anyone under 18, or anyone under the age of majority where that age is higher than the age of 18.

• create, advertise, solicit, distribute, or possess materials that sexually exploit minors, including child sexual abuse material (CSAM), such as AI-generated CSAM (AIG CSAM) and stylized, animated, or fictional depictions;
• sexualize or solicit the sexualization of any minor (real or fictional) in any medium, including in text, image, audio, or video, and including through roleplay and persona-based interactions;
• instruct on or assist with producing any of the material described above;
• create content that promotes, normalizes, or encourages the sexual abuse or exploitation of minors;
• groom, sexually solicit, sexually extort, or traffic a minor, or assist anyone else in doing so;
• expose a minor to age-inappropriate content, including sexual content, graphic violence, and graphic depictions of self-harm;
• promote to a minor unhealthy dieting or exercise, or dangerous challenges to a minor, or shame or stigmatize a minor’s body or appearance;
• unlawfully distribute information or materials to a minor, including obscene materials, or fail to apply age verification or age-gating that the law requires;
• enable a minor to access age-restricted goods, services, or activities;
• fail to report child sexual abuse material to the appropriate authorities where the law requires you to report it; or
• otherwise endanger or exploit a minor.

Endanger life or safety

• design, develop, acquire, or deploy chemical, biological, radiological, or nuclear weapons, high-yield explosives, or other weapons capable of causing mass casualties;
• unlawfully develop conventional weapons or firearms, or evade weapons control, sanctions, or export control requirements;
• control, operate, or otherwise manipulate critical infrastructure, including energy, water, transportation, and public safety systems, or heavy machinery, without qualified human control;
• create, distribute, or deploy malware, ransomware, or exploits, or conduct cyberattacks, against systems you do not own or are not authorized to test;
• unlawfully synthesize, acquire, or distribute drugs, or other prohibited or controlled substances;
• encourage, promote, or facilitate suicide, self-harm, or disordered eating, or generate graphic depictions of self-harm; or
• otherwise plan, develop, or carry out activities that create a serious risk of death, bodily harm, or destruction of property.

Violate privacy or conduct surveillance

• collect, process, disclose, generate, or infer private or sensitive information about a person, including about their identity, health, or demographic characteristics, without a lawful basis;
• track, monitor, or identify a person’s location, communications, or emotional state without their consent or lawful authority;
• create or expand facial recognition databases through the untargeted scraping of facial images from the internet or CCTV footage;
• engage in real-time remote biometric identification in publicly accessible spaces for the purpose of law enforcement, except as permitted by applicable law;
• use biometric categorization to deduce or infer a person’s race, political opinions, trade union membership, religious or philosophical beliefs, sex life, or sexual orientation;
• engage in emotion recognition in the workplace or in education, except where necessary for medical or safety reasons;
• engage in social scoring, meaning the evaluation or classification of people based on their social behavior or personal characteristics in a way that leads to detrimental or unfavorable treatment;
• engage in predictive policing or criminal risk assessment based solely on profiling a person or on assessing their personality traits and characteristics; or
• otherwise violate a person’s privacy or conduct unlawful surveillance.

Deceive or mislead

• commit fraud, run scams, phish, or engage in other deceptive financial or commercial practices;
• create, acquire, or distribute forged, counterfeit, or stolen credentials or identifiers;
• create or spread content intended to deceive people about facts, events, or its own origin, including harmful manipulation, or, except as permitted by law, generate defamatory content about a person or organization;
• send spam, or manufacture false online engagement, or create fake accounts or coordinated inauthentic activity;
• misrepresent your identity, qualifications, expertise, or affiliation, or another person’s;
• use a person’s name, likeness, voice, or persona without their consent or legal right to do so, including to impersonate them;
• remove, alter, obscure, or degrade any watermark, label, metadata, or other provenance signal that we apply to our models or their output, or distribute output with those signals stripped or falsified;
• deploy subliminal, manipulative, or deceptive techniques that materially distort a person’s behavior or impair their ability to make an informed decision; or
• otherwise intentionally deceive or mislead others.

Interfere with elections and democratic processes

• misrepresent the time, place, or manner of voting, the eligibility requirements for voting, or the outcome of an election;
• discourage or suppress lawful voting, including through intimidation or targeted demobilization campaigns;
• impersonate a candidate, an election official, a political party, or a government agency;
• generate or distribute political advertising or campaign material that conceals its sponsor, its origin, or the fact that it was generated by AI, where disclosure is required by law or necessary to prevent deception;
• simulate grassroots support or opposition, including by using multiple personas or accounts to make coordinated political messaging appear independent;
• interfere with the administration of an election, including with election infrastructure, voter registration systems, or the counting or certification of votes; or
• otherwise undermine the integrity of an election or a democratic process.

Act without authority

• access any system, network, account, or data without authorization, or beyond the authorization you have been given;
• direct an agent to take action on behalf of a person who has not authorized it;
• defeat CAPTCHAs, rate limits, robot exclusion protocols, or other measures that a third party uses to limit automated access to its systems;
• operate an agent without the practical ability to monitor, correct, or stop it;
• interact with people through an automated system in a way that conceals that the system or its outputs are automated, where a person would reasonably expect to be dealing with a human; or
• otherwise use automation to do indirectly what this policy prohibits doing directly.

Interfere with our products and services

The following apply to our products and services, including (a) activities you carry out without using them, and (b) third-party products and services incorporated into them. You agree that you will not:

• gain unauthorized access to our products and services, to our infrastructure, or to another user’s account, data, or workloads;
• probe, scan, or test the vulnerability of our systems, or breach or circumvent any security, authentication, or rate-limiting measure;
• create or distribute malicious code, malware, or computer viruses, or otherwise disable, overburden, interfere with, or impair the proper working, integrity, or operation of any of our computer systems;
• circumvent or remove usage restrictions or other safety measures, or enable functionality we have disabled;
• use our computing capacity for purposes unrelated to the products and services you obtained it for, including cryptocurrency mining; or
• otherwise interfere with our products and services, or with anyone else’s use of them.

Building responsibly

In addition to the specific prohibitions above, you agree that you will adhere to the following principles of responsible design and use :

• Not use our products and services to make, or to substantially automate, a consequential decision without meaningful human review. Consequential decisions are those that materially affect a person’s rights, safety, health, or access to essential services and opportunities, including decisions about employment, education, credit and lending, housing, insurance, healthcare, public benefits, and legal rights.
• If what you build carries material limitations you know about or foreseeable risks of harm, disclose them to the people who use it.
• If you make our products and services available to others, whether as a partner, as a reseller, or inside something you build, hold them to terms at least as protective as this policy.
• If you make our products and services available to minors, you are responsible for complying with the laws that apply, including laws on age assurance, privacy, content moderation, and disclosures.
• If your product engages people in open-ended or companion-style conversation, maintain a protocol for responding to users who express suicidal ideation or intent to self-harm, including referral to crisis services, and comply with laws that apply to companion or mental-health chatbots.

Reporting vulnerabilities

If you find a vulnerability in our products and services, or our infrastructure, report it to aup@reflection.ai.

Changes to this policy

We will update this policy as our products and the ways people use them change. When we do, we will post the revised version here with a new effective date.

Questions

Questions about this policy and requests for an exception may be sent to aup@reflection.ai.