AI use & transparency

An automated assistant, not a person

The Markaigen assistant is automated. Its current mode retrieves prepared knowledge-base answers; external generative-model answers are not enabled. It cannot access your private account data, browse live websites on your behalf, approve decisions or take administrative actions. An answer can be incomplete or outdated. Contact a person at Markaigen when accuracy, an account issue or a consequential decision needs attention.

Our editorial standard

AI may assist research organisation, drafting, translation, summarisation, code and visuals. The publisher remains responsible. Substantive public-interest content requires editorial responsibility and appropriate checking; a generic site-wide notice is not a substitute for a disclosure required on a particular item. Labels should distinguish AI-assisted work, generated material with human review and automated output. We do not present synthetic people, fabricated quotes or generated reviews as authentic evidence.

Tools, estimates and agents

Each tool must be read with its method and limits. A heuristic score is not an official search-platform score or compliance certificate. User-pasted answers describe the supplied sample, not market-wide visibility. Agent Studio stores workflow designs and an owner’s recorded approval; it does not execute autonomous agents or provide an independent second approver.

Data and provider changes

Do not enter secrets, sensitive personal information or unapproved client data. Current chat history is stored in your browser session; question logging and the external AI connection are disabled. Optional browser voice processing can involve the browser’s provider. See Privacy. A future provider connection requires an updated disclosure of provider, transmitted fields and relevant controls before activation; enabling a connection is not permission for unrelated model training.

Human responsibility and reporting

Check factual claims against primary sources, review bias and copyright risks, and assess the consequences of publishing or acting. Never use a generated output alone for decisions affecting employment, credit, health, legal rights or similarly significant interests. Report unsafe, misleading or undisclosed AI content through Corrections and complaints, with the page or tool and an example that avoids unnecessary personal data.

Our operational requirements include role-appropriate AI literacy, documented review and risk assessment before materially new AI uses. This policy sets that standard; it is not a claim that every existing item has been independently audited or that the platform holds an AI compliance certification.