Learn AI /Responsible AI, Human Rights and Safety /Transparency
Transparency
- Starter
- Responsible AI, Human Rights and Safety
- Policy officers
- Human rights practitioners
Also called: disclosure.
Back to the milestone you came from
In five seconds
What it is
Being open about where an AI system is used, what it does, what information it uses, and who remains responsible for its effects.
In plain words
In simple terms, transparency means not hiding the use of AI. People should be told that a tool is being used, what it helps with, what its limits are, and who is still responsible.
An everyday example
A parcel kiosk reads the postcode and measures the box before suggesting a delivery category. The screen says that an automated system made the suggestion, lists the information it used and warns that damaged labels can cause errors. A clerk checks rejected parcels, and the receipt explains how to correct an address or question a surcharge. The customer does not need the system's source code, but does need enough information to understand and challenge the decision. The postal service remains responsible for putting the parcel on the right route.
A library sign is simple and static, while real AI systems can change over time, use many data sources, and behave differently across contexts. Transparency in AI may need layered documents for different audiences, such as users, auditors, and affected communities. Also, a visible notice alone is not enough if the information is too vague, too technical, or impossible to act on.
Remember thisTransparency means disclosing where AI is used, what it does and relies on, its important limits, how people can question it, and who remains responsible.
This term is contested
The basic idea is widely accepted, but the exact amount and kind of disclosure required is context-dependent and sometimes debated, especially when transparency must be balanced with privacy, confidentiality, security, or intellectual property.
Understand it
Understand it
Why it matters
Transparency helps people recognise when AI may affect them and what questions to ask before relying on it. If a tool recommends, ranks, drafts, flags, or filters information, people need to know that it is being used, what it is meant to do, and where it may go wrong. That supports safer decisions, clearer accountability, and more meaningful human oversight. It also helps organisations spot unfairness, reduce blind trust in automation, and make it easier to challenge mistakes. Without transparency, people may not know that an AI system influenced an outcome at all, which makes review, remedy, and public trust much harder.
What transparency adds around an AI system
After viewing, the reader should be able to say that transparency means disclosing AI use, inputs, limits, and responsibility so people can question and oversee it.
01
1. AI is used
An organisation uses an AI system for a task such as ranking, recommending, or drafting.
Draw a box with a simple system icon and the label 'AI tool used for a task'.
02
2. Notice is given
People are told that AI is involved and what task it supports.
Draw an arrow to a box with a notice icon and short text: 'AI used here' and 'Purpose: recommend documents'.
03
3. Inputs and limits are stated
The organisation explains what information the system uses, where it may fail, and what it should not be used for.
Draw an arrow to a box with three bullet lines: 'Uses catalogue metadata', 'May miss recent items', 'Not the final decision'.
04
4. Human responsibility is clear
A named team or role checks outputs and remains accountable.
Draw an arrow to a box with a person icon and text: 'Human reviewer accountable'.
05
5. Questions and challenge are possible
Users and affected people can ask for clarification, correction, or review.
Draw an arrow to a final box with a speech bubble and loop-back line labelled 'Request review or correction'.
A left-to-right step diagram with five boxes. Read from first to last. The steps show an organisation moving from using an AI tool in secret to providing clear notice, useful documentation, human review, and a way to challenge results.
On hover or tap, each step reveals one sentence of plain-language detail and highlights the related part of the process.
Describe this without the diagram
The diagram is a five-step sequence read from left to right. Step 1 shows an organisation using an AI tool for a task. Step 2 adds a notice telling people that AI is involved and what it is meant to do. Step 3 adds information about what inputs the system uses and its limits, such as missing some items or not being suitable for final decisions. Step 4 shows that a human team remains responsible for checking outputs. Step 5 shows a way for people to ask questions, correct errors, or request review. The main message is that transparency surrounds the AI system with usable information and accountability.
With reduced motion: A static five-box sequence with arrows from left to right. All step details are visible at once, with no animation or hover-only content.
Transparency means giving people clear, usable information about an AI system. That includes saying that AI is involved, what task it is being used for, what information goes in, what kind of output comes out, and who to contact or challenge if there is a problem.
What goes in can include data, rules, instructions, or a prompt. What happens in the middle may be complex, but transparency does not require publishing every line of code. It requires enough disclosure for the right people to understand the system’s purpose, limits, risks, and oversight.
What comes out of transparency is not better predictions by itself. The output is information for people: notices, labels, documentation, model cards, impact assessments, public explanations, and routes for review. These help users, affected people, managers, auditors, and regulators ask informed questions.
Transparency is often confused with explainability and interpretability. Explainability focuses on explaining a specific output or behaviour. Interpretability usually refers to how understandable the internal workings are to specialists. Transparency is broader: it covers openness about the whole system, its use, its governance, and its limits.
How “The parcel kiosk receipt” maps onto the real thing
| In the story | In the AI system |
|---|---|
| The sign saying a software tool is being used | Disclosure that AI or automation is involved |
| The note listing catalogue information and known gaps | Information about inputs, scope, and limitations |
| The named staff member who checks results | Human oversight and accountability |
| The option to ask for review | A channel to contest or seek clarification |
A worked example
Before an AI recommendation feature is uploaded to a public knowledge portal, the web editor prepares the notice and supporting documentation that users will see beside suggested background documents.
Starting from: The feature uses public document metadata, including title, language, publication year, subject tags and publication status.
- The page states that the suggestions are generated by an AI tool using public document metadata.
- An information panel lists the input categories and confirms that the feature does not read private user files.
- The panel explains that newly uploaded or poorly tagged documents may be missed, while heavily tagged documents may appear more often.
- The web editor reviews a sample before release and records misleading matches, corrections and known limitations.
- Each recommendation page provides a route to report an error or request human review.
Result: Users can see where AI is involved, what information it uses, its main limitations, the oversight arrangement and how to question a result. The office remains responsible for the feature and for responding to reported problems.
In a United Nations workplace
Before close of business in a small field presence with unreliable connectivity, a humanitarian affairs officer conducts a records retention review of cleared situation reports for which an approved AI assistant proposed retention labels.
The note for the file records that AI was used, the categories of information examined, known limitations, any labels the officer changed, and the human clearance arrangement. The chief of section remains responsible for the retention decision, not the assistant.
Before you try thisReports containing personal, protection-sensitive, security-sensitive or operational information must not be entered into a tool unless it is approved for that information and purpose. Poor connectivity is not a reason to transfer files to a public service. Every proposed label requires human review because an incorrect retention decision may expose or prematurely delete records.
Where you may hear it
- At a vendor demonstration before a procurement unit opens the file for technical evaluation
- During an ICT governance board's first reading of an AI system proposal
- When a public notice for an AI-supported service goes through clearance
- During an internal audit review of system logs, oversight records and complaint routes
Avoid confusion
Avoid confusion
Often confused with
Explainability
Explainability means being able to show, in human-usable terms, why an AI system produced a particular output or pattern of outputs.
Interpretability
Interpretability is the extent to which people can understand how a model works or produces outputs, sometimes by inspecting its internal parts.
Side-by-side
Common misconceptions
Common misconception: Transparency means publishing the full model, code, and data to everyone.
Correction: Not always. Useful transparency depends on audience and risk. Public users may need clear notices and limits, while auditors or regulators may need deeper technical access. Full disclosure can conflict with privacy, confidentiality, security, or intellectual property, so the goal is meaningful information, not maximal exposure.
Common misconception: If a system is transparent, it is automatically fair and safe.
Correction: Transparency helps people inspect and question a system, but it does not remove bias, errors, or harmful uses by itself. An organisation can disclose many details and still deploy a system that performs badly or unfairly. Transparency supports accountability; it does not replace it.
Common misconception: A small label saying 'AI-generated' is enough.
Correction: A label is only a starting point. People often also need to know the system’s purpose, main inputs, limits, review process, and who is responsible. If the notice does not help someone act, challenge, or decide appropriately, it is not very meaningful transparency.
Quick check
Which example best shows meaningful transparency about an AI tool?
Meaningful transparency gives usable information, not just a label or a slogan. The best answer tells people that AI is used, what for, what it relies on in general terms, where it may fail, who reviews it, and how to question it.
Go deeper
Go deeper
The simple version
In simple terms, transparency means not hiding the use of AI. People should be told that a tool is being used, what it helps with, what its limits are, and who is still responsible.
The technical version
Technically, transparency is the practice of providing accessible disclosure and documentation about an AI system across its lifecycle. This can include notice of AI use, intended purpose, developers and deployers, data sources or input categories, performance information, known limitations, risk controls, human oversight arrangements, record-keeping, and routes for contestation or remedy. The exact level of disclosure may vary by role and context: end users, affected persons, oversight bodies, procuring organisations, and technical assessors may need different forms of information. Transparency is broader than explaining one output. It concerns system-level visibility, governance, and traceability, while also balancing privacy, confidentiality, security, and other legitimate constraints.
Reviewed external resources
Primary sourceRecommendation on the Ethics of Artificial Intelligence (opens in a new tab)UNESCO
This UNESCO recommendation gives an official international policy reference on ethical AI governance, and it suits policy, governance, and public-sector readers.
Intermediatelink checked 2026-07-31
Primary sourceB-Tech Project (opens in a new tab)UN Human Rights Office (OHCHR)
This OHCHR B-Tech Project resource helps readers connect transparency to business responsibility and human rights practice, and it suits human-rights, policy, and accountability audiences.
Intermediatelink checked 2026-07-31
EncyclopaediaWikipedia search: Transparency (opens in a new tab)Wikipedia
This Wikipedia search results page helps readers find several possible meanings and starting points for 'transparency', and it suits beginners who need orientation rather than one settled source.
search results, not one pagelink checked 2026-07-31
VideoVideo explanations: Transparency (opens in a new tab)YouTube search
This YouTube search results page helps visual learners find a range of video explanations on transparency, and it suits beginners who want different teaching styles.
search results, not one pagelink checked 2026-07-31
DiscussionCommunity discussion: Transparency (opens in a new tab)Reddit search
This Reddit search results page helps readers see how people discuss and question transparency in practice, and it suits readers who want informal examples and debate.
search results, not one pagelink checked 2026-07-31
Where this came from
Drafted with a language model through an offline pipeline, checked against this site's schema, and every external link above fetched to see whether it answers. Treat it as a starting point and check anything that matters against the primary sources. How this site is built and what that means.
The one thing to remember
Transparency means disclosing where AI is used, what it does and relies on, its important limits, how people can question it, and who remains responsible.