New Usage Policy: fifteen universal rules that apply to every model including RAI Own, extra care for high-risk uses such as legal, medical and financial advice, rules for agents acting on your behalf, and how enforcement and appeals work.
This Usage Policy is part of the Consumer Terms of Service of Aerendir Mobile Inc. and applies to every use of RAI: the web app, the Android apps, cloud models reached through our gateway, models that run on your device, and the "RAI Own" model we host ourselves. It describes what you may not do, the uses that need extra care, and what happens when the policy is broken. Where we say "you", we mean you and anything an agent, scheduled task or tool does on your behalf.
These standards apply to everyone, every model and every mode, including on-device and RAI Own.
Do not use RAI to do anything illegal where you are or where the effect is felt, or to help anyone else do so.
Do not use RAI to attack, disrupt or gain unauthorised access to power, water, transport, financial, healthcare, telecommunications or other critical systems.
Do not use RAI to create malware, to find or exploit vulnerabilities in systems you are not authorised to test, to steal credentials, to evade security controls, or to attack any system without the owner's permission.
Do not use RAI to design, develop, acquire or deploy weapons, including biological, chemical, nuclear or radiological weapons, explosives, or conventional weapons intended to cause mass harm.
Do not use RAI to promote, threaten, plan or glorify violence, terrorism or genocide, or to produce content that attacks people because of who they are.
Do not use RAI to collect, track, de-anonymise, dox or profile people without their consent, to impersonate a real person, or to forge identity documents, signatures or credentials.
Do not use RAI to create, request or distribute sexual content involving minors, to groom or exploit minors, or to expose minors to content that is harmful to them. We report child sexual abuse material to the authorities.
Do not use RAI to harass, bully, threaten or demean people, to encourage self-harm, eating disorders or suicide, or to manipulate people in ways that harm their wellbeing.
Do not use RAI to produce content you know to be false and intend to be believed as true, including fabricated news, fake reviews or doctored evidence, or to deny well-documented atrocities.
Do not use RAI to run coordinated inauthentic behaviour, sock puppets, fake engagement, astroturfing, or bots that pose as humans without disclosure.
Do not use RAI to interfere with elections, to suppress voting, to impersonate officials or candidates, or to spread false information about how to vote.
Do not use RAI for mass surveillance, biometric tracking, covert monitoring of individuals, censorship of lawful speech, predictive policing, or any law enforcement or criminal justice use that lacks lawful basis and human oversight.
Do not use RAI for scams, phishing, spam, pyramid schemes, deceptive advertising, plagiarism presented as original work, or cheating in academic or professional assessments where that is prohibited.
Do not share accounts, create accounts to evade a suspension or a limit, jailbreak or prompt-inject models to bypass safety behaviour, extract or distil models, scrape the service, overload it, or interfere with other users.
Do not use RAI to create sexual content depicting minors, real people without their consent, or non-consensual acts. Adult content involving fictional adults is permitted only on RAI Own and only within every other rule in this policy.
RAI Own is a model we host that applies fewer content filters than other models. It is available only to adults who have acknowledged, once per account, that it produces unfiltered output and that they accept sole responsibility for their prompts and for what it produces. Every universal standard above applies to RAI Own in full. Fewer filters is not permission: it means we rely on you, and we may still suspend accounts for violations we detect through reports, safety flags or abuse investigations, and may withdraw RAI Own entirely at any time.
Some uses are permitted but need extra care because an error can seriously affect someone's life, rights or safety. These include legal, medical and mental health, financial and insurance, employment and hiring, housing, credit, education and admissions, immigration, benefits, and control of physical devices or vehicles.
If you use RAI in one of these areas, you must:
General information and education are not high-risk use. A decision that is favourable to the affected person and gives them something they asked for does not need the human review above.
When RAI acts for you, through scheduled tasks, tools, email or device features, you are responsible for what it does as if you had done it yourself. Do not direct an agent to send bulk or unsolicited messages, to act on accounts or systems you are not authorised to use, to scrape or collect data in breach of a website's terms, to make purchases or commitments you have not approved, or to take any action listed above. Review an agent's work before relying on it.
We use automated systems and human review of reported or flagged content to find violations. If we believe you have broken this policy we may warn you, slow or limit your access, remove content, suspend or terminate your account, and, where the law requires, report to authorities. Being blocked by a safety system in the moment does not by itself mean you have broken a rule; patterns and reports do. If you think we got it wrong, write to legal@aerendir.ai with your account email and what happened, and we will look again. We may update this policy as described in the Consumer Terms of Service; every version is archived at https://aerendir.ai/legal/archive/.