AI Principles
Artificial intelligence offers new ways to support and improve how we work, but its use must never compromise the quality, safety, or trust at the heart of Psychology Tools.
Our AI Principles set out how we use AI responsibly: protecting the integrity of our clinical resources, safeguarding user and client privacy, and ensuring that people – not AI – remain accountable for the work we produce and the decisions we make.
Principle | What it means |
|---|---|
Human authored, expert reviewed. Our expertise is our reputation. Users should feel safe in knowing that the resources we provide are written and designed by humans, and reviewed by clinical experts. | • All published resources are human-authored, human-reviewed, and approved by qualified clinical and counselling psychologists and other mental health professionals. • Our resources are rigorously checked to ensure that they contain no oversimplified clinical information or unsafe wording or concepts. • We maintain a clear separation between human-authored clinical content generation, and any AI processes that are adopted. • Our resources are never forward-translated by AI, to avoid distorted or damaging clinical meanings. AI is sometimes used to back-translate resources into English to allow for translation quality checks. • We may use AI to generate internal outlines, drafts of marketing content and summaries of existing documents, or to edit, refine, or improve the clarity of text that has been written by human authors. • We may use AI to generate images and icons, where they are illustrative only and are not relied upon for clinical, medical, educational, or factual content. • We have strict internal guidance and processes on how AI services are to be used, ensuring that AI services are vetted and approved before use, that customer and client data is kept confidential and that all AI outputs are checked. |
Accountability, governance and transparency. AI may support our work, but it does not own our decisions. We remain accountable for how AI is used, where it is used, and how its outputs are reviewed. | • We produce clear customer-facing information on what AI is used for. • Guidance and training internally is consistently given on where AI use is permitted and prohibited. • Established regular reviews are made to ensure authorised use is still within ethical and practical guidelines. • We have clearly established risk levels for different AI use cases. • We have clear ownership and accountability roles, and named owners for AI adoption and use. |
User and client privacy. Trust depends on privacy. We protect user and client information by keeping sensitive data out of AI systems. | • We do not enter or upload identifiable data, mental health information, user submitted data, or any identifiable subscriber data to third party AI tools. • We only use vetted and approved AI vendors, services, and environments. • We continually assess third party AI vendors for risk, data usage, training practices, known biases, security flaws, and other issues. • We do not use AI in ways that could expose, infer, profile, or exploit subscriber behavior or user-related interests. |