RISOS AI Terms

1. AI-assisted functionality

RISOS uses model-assisted processing as one part of its decision-intelligence workflow. Depending on the research path, models may help extract facts, classify evidence, synthesize independent signals, verify a draft conclusion, or explain uncertainty. RISOS also uses deterministic rules, source-policy controls, evidence lineage, and other non-model processing.

2. Inputs and outputs

“Input” means prompts, research questions, instructions, evidence, or other material submitted or routed to an AI feature. “Output” means model-assisted text, classifications, summaries, assessments, recommendations, or other material returned by the feature. Input you submit is treated as your content under the Terms. RISOS does not claim ownership of your underlying content merely because it is processed by an AI feature.

AI output may not be unique, and similar material may be generated for other users. You are responsible for assessing third-party rights and the suitability of output for your intended use.

3. Accuracy, uncertainty, and human review

AI output may be incomplete, inaccurate, stale, misleading, or unsuitable for a particular purpose. RISOS is designed to preserve source evidence, uncertainty, and contradictions where available, but no confidence label or verification step is a guarantee. You remain responsible for reviewing evidence and exercising judgment before relying on a material result.

4. Professional and consequential decisions

RISOS is not a legal, medical, accounting, investment, credit, employment, insurance, housing, or other regulated professional adviser. Do not use AI output as the sole basis for a consequential decision affecting a person’s legal rights, eligibility, health, safety, employment, credit, housing, insurance, or similar interests.

5. Model providers and policy-controlled routing

RISOS may use OpenAI, Anthropic, xAI, local processing, or another approved model route depending on the feature, configured availability, and the policy that applies to the underlying evidence. Source-platform restrictions take priority over model convenience: when a source policy prohibits raw model processing or permits only a specified route, RISOS is designed to block or limit the incompatible route rather than treat unavailable permission as authorization.

The Privacy Policy describes the categories of information that may be processed and how service providers may receive information. Different model providers can have different contractual, retention, and operational terms; these AI Terms do not promise that every provider handles data identically.

6. Source-platform data

Google, YouTube, TikTok, X, and other source-platform data may be subject to provider-specific restrictions on model processing, storage, embeddings, derived metrics, export, or reuse. You may not use RISOS outputs, APIs, MCP tools, or another method to bypass those controls. Provider-specific terms in the Terms of Service and Privacy Policy remain controlling for affected data.

7. Sensitive information

Unless a feature expressly states otherwise, do not submit passwords, API secrets, payment-card data, government identifiers, regulated health records, children’s information, or other highly sensitive data to an AI feature. RISOS’s general research features are not designed for those data categories.

8. Actions and automation

RISOS may automate research, monitoring, retrieval, classification, and workspace operations within the permissions and limits of the service. A model-generated recommendation does not by itself authorize an external action, purchase, publication, deletion, or other consequential act. Where a connected tool or developer integration can take an action, the user or developer is responsible for the permissions and safeguards configured for that integration.

9. Transparency

RISOS identifies its service as model-assisted decision intelligence through its product, documentation, and these terms. Where applicable law requires an additional in-product AI disclosure, marking, or label for a particular interaction or generated content type, RISOS will apply the requirement to the affected feature rather than relying on these terms alone.

10. Prohibited AI uses

You may not use AI features to violate law or rights, deceptively impersonate another person, remove a legally required AI disclosure or provenance signal, bypass safety or access controls, process restricted data contrary to product instructions, or engage in activity prohibited by the Acceptable Use Policy.

11. Model and feature changes

Models, providers, routing, and AI features may change. A material change that affects data handling, source-policy compatibility, retention, user rights, or legally required transparency is subject to legal and product review before the new use is represented as available.

12. Contact

Questions about RISOS AI features can be sent to hello@risos.co.