presentsUNAIVERSE
The UN AI timeline

    Learn AI /Generative Media /Misinformation and disinformation

    Misinformation and disinformation

    Also called: disinformation, misinformation.

    Detail

    In five seconds

    What it is

    False or misleading content can spread by mistake or on purpose, and the difference matters because intent changes how we respond.

    In plain words

    Misinformation is wrong content shared by people who think it is true. Disinformation is wrong content shared to fool people on purpose.

    An everyday example

    In a school trip group chat, Amina writes, “Samir said he cannot come because he is sick.” Several people are unsure whether “he” means Samir or the bus driver mentioned earlier. Luis guesses it means the driver and tells others the trip might be cancelled. Later, one student who wants the trip postponed repeats the same claim even after seeing the teacher’s correction. The same unclear message led first to an honest mistake and then to deliberate spreading of a false claim.

    Where the story stops working

    This analogy uses ambiguity in one message, but real misinformation and disinformation often involve more than confusion. They can include fabricated images, synthetic audio, coordinated campaigns, advertising systems, and repeated reposting across many platforms. Also, in real cases, a message may be partly true rather than simply misread.

    Remember thisThe same false claim can be an honest mistake or a deliberate deception, and that difference shapes the safest response.

    This term is contested

    Definitions are widely used but not perfectly uniform across organisations. The usual distinction is intent, yet some institutions classify related terms such as malinformation differently or apply the labels with different thresholds of evidence.

    Understand it

    Understand it

    Why it matters

    People often need to act before all facts are clear. False or misleading content can affect health choices, voting behaviour, reputations, safety decisions and trust in institutions. The distinction between misinformation and disinformation helps you choose a proportionate response. An honest mistake may call for clarification and better context. A deliberate campaign may require monitoring, evidence gathering, platform escalation, and stronger risk communication. Generative AI matters here because it can make convincing text, voices and images quickly and cheaply, which can increase volume and make checking slower. But AI is only part of the problem; old-fashioned rumours and edited media still spread widely.

    Same false claim, different intent

    Show that misinformation and disinformation can look similar on the surface, but differ in whether the sharer aims to mislead.

    1. 01

      1. False or misleading content appears

      Both paths begin with a claim, image, audio clip or video that is wrong, manipulated or lacking important context.

      Top row spanning both columns: one source card labelled “False or misleading content” with small icons for text, image, audio and video.

    2. 02

      2. Why it is shared

      In misinformation, a person shares it because they think it is true. In disinformation, a person or group shares it to mislead others.

      Second row split into two columns. Left card: “Believed to be true”. Right card: “Shared to mislead”. Use different colours and simple arrow connectors.

    3. 03

      3. Spread

      People repost, quote, translate or clip the content. Recommendation systems and group chats can increase its reach whether the original sharing was careless or deliberate.

      Third row in both columns: branching arrows from one card to several smaller audience cards. Add a small note under both: “Spreads quickly”.

    4. 04

      4. Effects and response

      Both can cause confusion or harm. A useful response is to verify before sharing, add context, correct clearly, and investigate intent when that matters.

      Bottom row spanning both columns: one outcomes card labelled “Confusion, false beliefs, harm”. Beneath it, a response card with four bullet icons: verify, contextualise, correct, investigate intent.

    A two-column comparison shows one false claim travelling through parallel paths. The left path is an honest mistake; the right path is deliberate deception. Read from top to bottom, then compare the outcomes and response needed in each column.

    On hover or tap, each column highlights its defining feature: left highlights “no intent to mislead”; right highlights “intent to mislead”.

    Describe this without the diagram

    The diagram compares two paths for the same false claim. First, false or misleading content appears. Then the paths split. In the left column, the content is shared because the person believes it is true. That is misinformation. In the right column, the content is shared to mislead others. That is disinformation. Both paths then spread through reposts and recommendations. Both can lead to confusion, false beliefs and harm. The recommended response in both cases is to verify, add context and correct, while also examining intent when relevant.

    With reduced motion: A static two-column chart replaces any highlighting. The left column reads: false content, believed true, spread, harm. The right column reads: false content, shared to mislead, spread, harm. A final note below both columns says: verify, contextualise, correct, and investigate intent.

    These terms describe information problems, not a single technology. Misinformation is false or misleading content shared without an intent to mislead. Disinformation is false or misleading content created or shared with the aim of misleading people. Both can appear as text, images, audio or video, including content made with generative AI.

    What goes in can be a rumour, a copied post, an edited image, a synthetic voice clip, or a video taken out of context. What happens next is that the content is created, altered, reposted or recommended, often moving faster than checking. What comes out is confusion, a false belief, or a damaged reputation, and sometimes real-world harm.

    The key difference is intent. If someone forwards the wrong evacuation time because they believed it, that is misinformation. If someone fabricates the time to cause panic, that is disinformation. In practice, intent can be hard to prove, so people sometimes use the broader phrase “misinformation and disinformation” together.

    These terms are often confused with deepfakes and hallucinations. A deepfake is one possible form of synthetic media, often audio or video, that may be used for misinformation or disinformation but is not the whole problem. A hallucination is a model output that is false or unsupported; it becomes misinformation only if someone shares it as if it were true, and disinformation only if it is used deliberately to mislead.

    How “A confusing message in a group chat” maps onto the real thing

    In the storyIn the AI system
    An ambiguous message with an unclear “he”A piece of content that is easy to misread or take out of context
    Luis repeating the wrong meaning because he guessedMisinformation shared without intent to mislead
    The student repeating the false claim after seeing the correctionDisinformation spread with intent to mislead
    The teacher’s correction arriving lateFact-checking often moves more slowly than sharing

    A worked example

    A synthetic voice message, made to resemble a community centre manager, falsely says that an after-school programme has been cancelled.

    Starting from: Voice message: “There will be no programme next week. Please make other arrangements.”

    1. The creator generates the imitation while knowing that the programme is continuing, then posts it with the centre's name.
    2. A parent believes the message is genuine and forwards it to other families as a warning.
    3. The centre publishes the current timetable and confirms that the voice message is not authentic.
    4. Reviewers preserve relevant public evidence and examine how the message was created and distributed rather than inferring intent from falsity alone.

    Result: The creator's message is disinformation because it was produced and distributed with intent to mislead. The parent's forward is misinformation because the parent believed it was true. The content is the same, but the intent of each person sharing it differs.

    In a United Nations workplace

    An evaluation officer in an inter-agency secretariat is clearing a side-event concept note when a graphic begins circulating that claims a draft evaluation found the initiative had failed. The graphic reproduces one chart but removes the qualification beneath it.

    Before close of business, the officer must establish whether the qualification was dropped through careless editing or removed to mislead. That distinction informs the response: clarify the finding and correct the concept note, or also document the distribution pattern and refer the matter to communications and information integrity colleagues.

    Before you try thisDo not upload interview notes, respondent identities, raw evaluation data or partner-sensitive findings to an external AI service that has not been approved for that material. Doing so can breach confidentiality and expose respondents to re-identification. Work from cleared, minimised extracts in approved systems, with human review.

    Where you may hear it

    • During clearance of talking points after a false quotation begins circulating
    • At an inter-agency working group reviewing an information integrity incident
    • In the front office before a public correction is issued
    • During a vendor demonstration of synthetic-media detection tools

    Avoid confusion

    Avoid confusion

    Often confused with

    Common misconceptions

    • Anything false online is disinformation.

      Not necessarily. Disinformation involves an aim to mislead. Many false posts are misinformation because the person sharing them thinks they are helping, warning others or repeating something they believe. The content may look the same, so intent has to be assessed carefully.

    • If AI made the image or voice, it must be disinformation.

      AI-generated content can be used for art, accessibility, training or harmless parody. It becomes misinformation or disinformation only when false or misleading content is presented in a way that causes people to be misled. The method of creation is not the whole definition.

    • You can always tell the difference by looking at the content.

      Usually you cannot. The same image, clip or sentence can be shared by one person who is mistaken and by another who is trying to deceive. To judge disinformation, you often need context such as patterns of behaviour, prior correction, coordination, or evidence of intent.

    Quick check

    Which statement best distinguishes misinformation from disinformation?

    Go deeper

    Go deeper

    The simple version

    Misinformation is wrong content shared by people who think it is true. Disinformation is wrong content shared to fool people on purpose.

    The technical version

    In information integrity work, misinformation and disinformation are related but distinct categories of false or misleading content. The usual distinction is intent: misinformation lacks intent to mislead, while disinformation includes it. The content may be fabricated, manipulated, wrongly attributed, or presented without crucial context. Distribution can occur through interpersonal sharing, broadcast media, search ranking and recommender systems. Generative AI can lower the cost of producing persuasive synthetic media, but these categories are defined by the content’s falsity or misleading framing and, for disinformation, the actor’s purpose, not by any specific model or platform. The boundary can be difficult in practice because intent is inferred from evidence, not read directly from the content itself.

    Reviewed external resources

    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

    The same false claim can be an honest mistake or a deliberate deception, and that difference shapes the safest response.