AI Use Disclosure

AI is optional

AI features are not required for manual coding, analysis, Results Builder, exports, or ordinary workspace use. No research material is sent to an AI provider unless you intentionally start an AI feature.

What may be sent when you use AI

The exact scope depends on the task. It can include selected text, a document or interview-answer scope, codebook labels and definitions, relevant memos, project description and research context, metadata summaries, selected evidence, and visual or table data needed to prepare the requested output. For transcription, audio from the selected media is processed to create the requested transcript. The workspace uses the OpenAI API for these features.

The assistant is designed to use the scope you select and does not independently browse unrelated projects. You should still avoid sending direct identifiers, confidential material, or content for which you do not have an appropriate right to process with an external AI provider.

What AI can and cannot do

AI can propose codes, code hierarchies, question-organised comparisons, project guidance, evidence-grounded answers, transcripts, results text, titles, and interpretations. Suggestions are provisional. You decide whether to accept, reject, edit, adjust evidence, create codes, finalize coding, or include material in reporting.

AI output can be incomplete, inaccurate, biased, or inappropriate for your methodological approach. It is not a substitute for research judgement, participant safeguarding, ethics review, legal advice, statistical testing, or professional interpretation.

Your responsibilities

• Review every suggested code, quotation, transcript, interpretation, and citation before relying on it.
• Check quotations against their source and protect negative, ambiguous, and minority evidence rather than using AI output as a conclusion.
• Use de-identified material where your study, consent materials, ethics approval, or institutional policy require it.
• Disclose AI use in your research reporting where your institution, funder, publisher, or methodology requires it.

Languages and data minimisation

The interface language and AI output language may be set separately. Choose the smallest useful scope for each task and use the evidence-selection controls to keep AI context proportionate to the work you are doing.

Questions

For questions about an AI task or a privacy concern, use Contact Support. For more information about personal-data processing, read the Privacy Notice.