People are more than data points.
Ask how an AI-assisted decision affects a person: what information it uses, who is responsible and how someone can raise an error and obtain a meaningful review.
Principles & governance · Edited by PsyBorg · Updated 2026-11-16
Start with the person affected
When an AI system helps an organization make a decision, ask how a person experiences the result. Can they understand what happened, identify an error and reach someone who can investigate it?
UNESCO's Recommendation on the Ethics of Artificial Intelligence, adopted in November 2021, places human rights and dignity at its center. UNESCO's official overview explains principles including privacy, fairness, transparency, accountability and human oversight. [1]
PSYBORG's practical questions below focus on turning those concerns into a process someone can use. They are editorial recommendations, not a statement that every question corresponds to the same legal entitlement in every country.
Ask what information enters the decision
Before considering a system's explanation, identify the information it uses and its role in the workflow. Does it summarize an application, recommend an outcome or carry out the final action? Who decides how to use its output?
PSYBORG suggests asking the organization to explain what information is needed for that purpose, who can access it and how long it is kept. Also ask how incorrect information can be identified and handled. Do not collect extra personal information simply to make a review form more detailed.
UNESCO's overview describes privacy protection throughout the AI lifecycle and emphasizes that human responsibility remains central. These principles help frame the questions; they do not establish what a particular organization has implemented. [1]
A hypothetical example: an incomplete application summary
Imagine a community training programme using AI to summarize applications for staff. An applicant receives an unsuitable timetable after a line about their availability is omitted from the summary.
An explanation that “AI helped process applications” would leave the practical problem unresolved. In PSYBORG's suggested process, the applicant could point out the missing information, receive acknowledgement and have a responsible person examine the original application and the resulting decision.
The organization would also consider whether similar omissions affected other applications and whether its workflow needs to change. This scenario is hypothetical. It makes no claim about a real programme or an automatic right to a particular outcome.
Make the review route usable
PSYBORG recommends checking the route from the affected person's perspective, including people who may find a complex form or digital-only process difficult to use.
- Contact: can the person find an understandable way to raise a concern without having to identify the model or technical fault?
- Issue: can they describe the outcome and the information they believe is wrong, while sharing only what is needed?
- Responsibility: does someone with authority to investigate receive the concern and know what action they can take?
- Response: is the next step explained, including how progress will be communicated and what happens if further review is needed?
- Follow-through: if an error is found, does the organization check the underlying information and any affected decision or record, then explain what it did?
Test the route with a hypothetical case before relying on it. Observe whether the concern reaches the right person and whether the response addresses the actual issue. A successful form submission shows only that the form accepted an entry.
Separate a principle from an assurance
A policy statement can name privacy or fairness while leaving practical questions unanswered. PSYBORG suggests looking for specific responsibilities, understandable information and evidence that concerns receive an appropriate response.
The right response depends on the system, its consequences and the people affected. Applicable rights and formal complaint routes also depend on the jurisdiction and circumstances. Use the relevant official rules and qualified advice when a specific legal decision is involved.
Sources and scope
[1] UNESCO, Recommendation on the Ethics of Artificial Intelligence. Adopted November 2021. Official overview: Core values and Core principles, particularly privacy and data protection, responsibility and accountability, transparency and explainability, human oversight, and fairness and non-discrimination.
Read source (www.unesco.org)
Official overview checked 16 November 2026. The application scenario, five review questions and suggested exercise are PSYBORG editorial analysis. They provide a starting point for examining a workflow, not a complete rights assessment or evidence that a particular system meets the Recommendation.
Continue reading
Privacy in AI deployments
Read source (psyborg369.com)
Responsibility is a process, not a promise.
Read source (psyborg369.com)
READING ROOM ↗