AI Transparency & Limitations
Opis uses generative AI as part of its research, interpretation and content-generation workflows. AI systems can make mistakes. This page explains what that means in practice and the controls Opis provides to help you inspect the work.
What AI does inside Opis
Opis uses AI to interpret unstructured information, assist with research, extract findings, explain results, generate or adapt content, and create images. Depending on your settings, the reasoning model may be Gemini or another model you configure.
AI is one component of the system. Where an Opis task uses deterministic scoring rules, the extracted observations and the calculation layer are kept conceptually separate so that a language model is not asked to improvise the mathematics.
Generated outputs are not Opis opinions
Opis retrieves, organises and analyses source material and user-supplied context to produce research and content outputs. Those outputs do not represent the personal views, beliefs or endorsements of Opis, its owner or its operators.
Where an output contains a recommendation, assessment or judgment, it is a generated analytical result based on the available evidence, configured methodology and supplied context. It should not be interpreted as a statement of Opis corporate opinion or endorsement.
Source material can itself contain opinions, claims or bias. Retrieving, citing, summarising or analysing that material does not mean Opis adopts or endorses the views expressed by the source.
Accuracy, confidence and uncertainty
A fluent answer is not automatically a correct answer. Sources can be wrong or stale, relevant evidence may be unavailable, and an AI model can misunderstand context or generate an unsupported statement.
Opis therefore surfaces confidence, sources and uncertainty signals such as assumptions, gaps, ambiguities and contradictions where the workflow supports them. A confidence label is an aid to review, not a guarantee of accuracy.
Evidence and human correction
Research outputs are designed to keep supporting sources inspectable rather than hiding them behind a final answer. Where Interactive Contextual Correction is available, you can add relevant real-world context, clarify an assumption or gap, and run a revised research pass.
Human correction improves the available context; it does not make the resulting AI output infallible. The revised result should still be reviewed in proportion to the importance of the decision.
Generated content and images
Content Studio and Image Studio can produce AI-generated or AI-assisted material. Review factual claims, citations, names, numbers, visual details, intellectual-property considerations and destination-platform requirements before publication.
Uploading or describing a visual reference provides context for generation; it does not guarantee an exact, factual or rights-cleared reproduction.
Consequential decisions
Opis is designed as a research and decision-support workspace, not as the sole decision-maker for matters that determine a person’s legal rights, employment, credit, access to essential services, medical treatment or other similarly significant outcomes.
For regulated, safety-critical or otherwise high-impact use, obtain appropriate professional review and apply the legal and governance requirements that apply to your organisation and jurisdiction.
Your role when using an output
Use the evidence, uncertainty indicators and editing tools provided by Opis. The more consequential the use, the stronger the verification should be. If a source does not support a material claim, or a result conflicts with reliable information you already hold, investigate it before relying on the output.