Welcome everyone to this course on critical multimodal thinking. Today we will examine how images, language, sound, editing, layout, and digital interactions work together to influence what people think, feel, and share. The goal is not to become suspicious of everything or cynical about media. The goal is to become more observant and make better-informed choices. We will decode real media, check questionable claims, discuss ethical communication, and develop activities you can use with young people. This course is supported by Erasmus+, but the views expressed belong to the course authors and participants.

This is a practical and interactive course designed for youth workers. We will explore how multimedia influences young people and how we can help them respond thoughtfully. The session should take approximately three to four hours, depending on how much time we give to activities and discussion. We will alternate between short explanations, group exercises, reflection, and content creation. By the end, you should have several methods you can reuse in workshops or informal conversations with young people. Please bring your own experiences and examples into the discussion while respecting privacy and avoiding identifiable sensitive information.

The course follows four connected stages. First, we notice how media attracts attention and produces emotional reactions. Second, we decode the choices behind a message: who created it, what it encourages us to believe or do, and which techniques create its effect. Third, we create our own messages while considering evidence, consent, privacy, inclusion, and potential harm. Finally, we apply these ideas to youth work through activities, difficult conversations, and safer digital spaces. The aim is to move from automatic reaction toward intentional interpretation and responsible participation.

Synthetic media is visual, audio, video, or written content that has been created or substantially altered using artificial intelligence or related digital techniques. Synthetic media includes: AI-generated images Cloned voices Deepfake videos AI-generated articles or comments Virtual influencers Altered faces or expressions AI-generated music and animation Synthetic media is not automatically deceptive. It can be used transparently for creativity, education, accessibility, translation, entertainment, and privacy protection. It becomes problematic when people are misled about what is authentic or when someone’s identity is used without consent. Example: AI-generated disaster image An image appears to show a dramatic flood in a European city. It is shared with the caption: “The media are hiding what happened yesterday.” The image was generated with AI, but it is reposted as documentation of a real event. Do not rely only on visual clues such as strange fingers or distorted signs. Instead, ask: Who first published the image? Is there an identifiable photographer? Do credible local sources report the event? Can the location and date be confirmed? Is there an earlier version with a different caption? Example: cloned voice A youth worker receives a voice message that sounds like a colleague: “I’m in trouble and need you to transfer money immediately. Please don’t call because I can’t speak safely.” The voice may have been cloned from publicly available recordings. The urgency and request for secrecy are warning signs. The recipient should verify the request through a separate, previously trusted channel. Example: deepfake video A video appears to show a school principal making an offensive statement. Before reacting or sharing it, viewers should seek: The original complete recording Confirmation from reliable sources Evidence about when and where it was recorded Other recordings of the same event A statement from the people or institution involved Example: synthetic persona A social-media account appears to show a young lifestyle influencer. The person’s photographs, biography, voice, and posts are generated or heavily managed using AI. Followers may not realize that the “person” does not exist. Questions include: Is the artificial nature of the account disclosed? Who operates it? What products or ideas is it promoting? Is it encouraging emotional attachment under false pretences? Is it imitating a real person without permission?

4. Bias Bias is a tendency or perspective that influences which information we notice, select, believe, or present. Everyone has biases, including journalists, educators, researchers, influencers, youth workers, and audiences. Bias does not automatically mean that information is false. It means we should examine how perspective and selection have shaped it. Example: confirmation bias A person believes that social media is destroying young people’s mental health. They notice and share every story about online harm but ignore research showing mixed results or positive experiences. Another person believes social media is entirely beneficial. They highlight connection and creativity while ignoring harassment, comparison, or compulsive use. Both people may use genuine examples, but they select the evidence that supports what they already believe. Example: selection bias in a video A creator interviews 20 people about immigration. Eighteen give moderate answers, but the video includes only the two most angry responses. The people in the video really made those statements. However, the selection creates a misleading impression of the whole group. Example: visual bias A news report consistently uses unflattering photographs of one politician and professional photographs of another. Even before reading the article, viewers may receive different emotional impressions. Questions for participants: Which information was selected? What may have been left out? Does the evidence represent the wider situation? Am I accepting this easily because it agrees with me? What evidence might change my conclusion?

Purpose: What grabbed your attention today? Time: 8 min. Facilitation: Read the steps aloud. Keep groups focused on observable features rather than guessing motives. Debrief by asking what changed their interpretation and how they could use the activity with young people. Let us begin with our own attention. Think back to the first online item that caught your attention today. It could have been a video, notification, headline, photograph, advertisement, or message. In pairs, describe what made you stop and look. Was it a face, a surprising statement, movement, music, color, conflict, or something personally relevant? Then identify the emotion it produced. Try to separate what the item said from how its design affected you. We are not judging whether clicking was good or bad; we are learning to observe how attention is guided.

Ask for one example from participants’ practice for each theme. Avoid blaming young people or platforms; emphasize that all people are vulnerable to design, repetition and social proof. Youth work now takes place inside an attention economy. Platforms, advertisers, creators, organizations, and political groups compete for limited attention. Their messages affect more than purchasing decisions. Repetition and selective framing can shape what feels true. Influencers, memes, and group signals can affect identity and belonging. Comparison, outrage, and constant exposure to distressing content can affect well-being. This does not mean young people are passive or uniquely vulnerable; adults respond to the same mechanisms. Our task is to help young people recognize influence while preserving their agency, curiosity, and enjoyment of media.

Example: reporting on a youth demonstration Two news outlets report on the same demonstration. Version A: “Young people demand urgent action on climate change.” The photograph shows young participants holding creative signs and listening to a speaker. Version B: “Protesters disrupt traffic in city centre.” The photograph shows an angry-looking participant standing near a blocked road. Both reports may contain true information, but their frames are different: Version A emphasizes participation and political engagement. Version B emphasizes disruption and public inconvenience. Neither image necessarily represents the entire event.

Questions for participants: What does each version encourage us to notice? Which people are shown and quoted? What information is missing? How does the photograph affect the headline? Could both versions be factually accurate but incomplete? Framing is not necessarily dishonest. Every message requires choices. Critical thinking means recognizing those choices and looking for relevant context.

Stress that no mode is neutral in context. The same sentence can feel trustworthy, comic, urgent or threatening when other modes change. A multimodal message creates meaning through several modes at once. We see images, colors, facial expressions, camera angles, and layout. We hear voices, music, rhythm, silence, and sound effects. We read captions, labels, hashtags, and emojis. We also experience pacing, cuts, movement, proximity, and interactive cues. None of these elements works entirely alone. A neutral sentence can feel trustworthy, threatening, humorous, or urgent depending on the accompanying image, voice, and editing. Critical multimodal thinking therefore asks not only, “What does this message say?” but also, “How does the complete experience shape its meaning?”

Purpose: Decode a short-form post Time: 12 min. Facilitation: Read the steps aloud. Keep groups focused on observable features rather than guessing motives. Debrief by asking what changed their interpretation and how they could use the activity with young people. Choose a recent social-media post or use a screenshot supplied by the trainer. Begin by listing everything you can observe: words, images, colors, expressions, camera position, sound, music, editing, emojis, and interaction cues. Avoid guessing the creator’s intention at first. Next, identify which mode carries the strongest emotion. Imagine removing that element—for example, muting the music or replacing the photograph. How would the meaning change? During the debrief, explain which feature most influenced your interpretation and how this exercise could be adapted for young people.

Decode a short-form post

Work in groups. For each post, identify every mode you can detect or infer. Because these are static simulations, imagine the sound/editing described in the caption. Purpose Notice how words, visual style, sound cues and platform signals combine to create meaning. | Suggested time: 12–18 minutes Work in groups. For each post, identify every mode you can detect or infer. Because these are static simulations, imagine the sound/editing described in the caption. EXAMPLES OF SIMPLE POSTS YOU CAN MAKE

Demonstration option: say the same neutral sentence—‘The meeting starts at six’—in three styles. Participants identify what changed even though the literal content did not. Style influences interpretation before we have evaluated the facts. Muted colors, stable framing, and a measured voice can make a message appear serious or authoritative. Fast cuts, playful sounds, and exaggeration can lower our resistance because the content feels entertaining. Tension music, close-ups, and urgent text can make the same information feel more dangerous or consequential. These signals do not automatically make a message true or false. They influence how we approach it. Consider which version you would trust, remember, or share, and then ask whether that reaction comes from evidence or presentation style.

This is the central decoding checklist. Ask participants to photograph this slide. In youth work, phrase questions with curiosity so they open conversation instead of testing young people. These five questions provide a simple method for slowing down interpretation. First, who created or shared the message, and what do we know about that source? Second, why was it made—what should we think, feel, buy, or do? Third, how do its images, language, sound, and editing produce that effect? Fourth, what is missing, including absent voices, evidence, history, or uncertainty? Finally, who benefits through money, attention, status, influence, or belonging? These are exploratory questions, not a test with one correct answer. Use them to open a conversation rather than interrogate young people.

Explain the loop without claiming that every platform or creator has the same intent. The practical point is to notice when emotion outruns evaluation. Many attention systems reward emotional intensity because strong reactions can produce more viewing, commenting, and sharing. The pattern can become: stop, feel, react, and repeat. Fear and outrage create urgency and narrow our attention. Identity cues turn content into a test of whether we belong to a group. Likes, novelty, and unpredictable rewards encourage us to continue scrolling. This does not prove that every platform or creator intends to manipulate people. The practical lesson is that emotional speed can overtake careful evaluation. When we notice a strong reaction, we can pause before acting on it.

Purpose: Turn emotional design into observable evidence Time: 18 min. Facilitation: Read the steps aloud. Keep groups focused on observable features rather than guessing motives. Debrief by asking what changed their interpretation and how they could use the activity with young people. In groups, examine a headline, thumbnail, meme, advertisement, or influencer post. Mark the element designed to create the strongest emotional response. Name the technique as precisely as possible: urgency, fear, social proof, identity, omission, exaggeration, or another device. Support your interpretation with something observable rather than guessing the creator’s motives. Then rewrite or redesign the message so that it communicates the essential information without unnecessary manipulation. In the debrief, compare what was lost and gained. Ethical communication does not need to be emotionless, but the emotional effect should be proportionate and honest.

Spot the manipulation

Purpose Turn emotional design into observable evidence and practise a proportionate rewrite. | Suggested time: 18–22 minutes

For each item: mark the emotional trigger, name the technique, identify the checkable claim, and rewrite the message to inform without unnecessary pressure.

Use a real example only if it is suitable for the group. Ask: What is being sold besides the product—an identity, lifestyle, value or relationship? Avoid shaming young people for enjoying creator content. Influencer communication can feel powerful because promotion is woven into a personal relationship. Repeated exposure may create parasocial closeness—a one-sided sense that we know and trust someone. Casual speech, confessions, and behind-the-scenes content can make a carefully managed performance feel spontaneous. At the same time, sponsorships, affiliate links, and self-promotion may be embedded in everyday stories. Ask what is being sold beyond the product. It could be an identity, lifestyle, value, or sense of belonging. Avoid shaming young people for following creators. Help them enjoy the content while recognizing the commercial relationship.

A parasocial relationship is a one-sided feeling of familiarity, friendship, or emotional connection with a public figure, influencer, streamer, celebrity, or fictional character. The audience may know many personal details about the creator, but the creator usually knows little or nothing about individual followers. The relationship can still feel genuine because followers regularly see the creator speaking directly to the camera, sharing personal experiences, or interacting with comments. Parasocial relationships are not automatically harmful. They can provide entertainment, inspiration, learning, and a sense of community. The concern arises when perceived friendship is used to influence trust, spending, identity, or behaviour. Example: influencer marketing An influencer regularly shares videos about anxiety, friendships, and daily life. Followers feel that the influencer is honest and relatable. The influencer then says: “You know I would never recommend something I don’t personally believe in. This supplement completely changed my life.” The video may feel like advice from a trusted friend, but it is sponsored and the influencer receives money or commission. The commercial message is strengthened by: Direct eye contact with the camera Personal confession Informal language Repeated exposure Claims of honesty and friendship Limited-time discount codes Questions for participants: What makes this person feel trustworthy? Is the commercial relationship clearly disclosed? Which claims can be independently checked? Would we trust the recommendation as much if it came from a traditional advertisement? What does the creator know about the individual follower’s needs? The objective is not to shame young people for following influencers. It is to help them distinguish emotional closeness from independent expertise.

Discussion: When does humor become harmful? Look for impact, power, repetition, target, context and whether people can safely opt out. Humor can carry ideas past our normal critical filters because we are focused on understanding and enjoying the joke. Memes compress complicated situations into a familiar image and a short piece of text. Irony can provide plausible deniability: if challenged, someone may say the audience took the joke too seriously. Shared references also create community, but they can define who belongs and who becomes the target. “It is only a joke” does not automatically erase harm. When evaluating humor, consider its impact, the power relationship, repetition, context, target, and whether people can safely object or opt out.

Purpose: Analyze a meme without killing the humor Time: 15 min. Facilitation: Read the steps aloud. Keep groups focused on observable features rather than guessing motives. Debrief by asking what changed their interpretation and how they could use the activity with young people. Select a meme and begin by describing only what is literally visible. Then identify the background knowledge someone needs to understand the joke. Ask what belief, assumption, or stereotype makes the joke work. Consider who is included in the shared understanding, who is excluded, and who may be positioned as the target. Finally, create an alternative version that keeps the humor while reducing stereotyping or harm. The purpose is not to destroy humor or find offence everywhere. It is to understand how humor carries meaning and to explore whether creativity can produce a joke without relying on someone’s humiliation.

Follow the joke to its message

Purpose Analyze humor without assuming that either amusement or offence settles the meaning. | Suggested time: 15–20 minutes Describe the literal image and text, identify the shared knowledge needed to understand it, then consider implied beliefs, power and impact.

Clarify: bias does not automatically mean a claim is false. It means perspective and selection shape the representation. Critical thinking asks what evidence supports the frame and what alternatives exist. Bias often appears through selection rather than complete fabrication. A headline tells us what aspect of an event to notice. A photograph can make someone appear powerful, vulnerable, aggressive, or isolated. Editing and sequence can suggest a causal relationship that the full recording does not support. Labels and metaphors can assign blame, innocence, status, or threat. Missing history, uncertainty, or relevant voices also changes the representation. Bias does not automatically mean that a claim is false. It means that perspective and selection shape what we see. Critical thinking asks what supports the frame and what reasonable alternatives exist.

Purpose: Find the frame before choosing a side Time: 18 min. Facilitation: Read the steps aloud. Keep groups focused on observable features rather than guessing motives. Debrief by asking what changed their interpretation and how they could use the activity with young people. Compare two reports, headlines, or posts about the same event. Underline words that carry judgment or communicate more certainty than the evidence supports. Compare the photographs, quoted sources, order of information, and missing context. What does each version encourage the audience to notice first? Then write a neutral headline that preserves the verified facts without pretending that every interpretation is equally supported. Finish by naming one question that remains unanswered. The goal is not to choose a side as quickly as possible. It is to recognize the frame before deciding what conclusion the available evidence justifies.

Compare two reports about one event Purpose Recognize framing through selection, language, images and omissions. | Suggested time: 18–25 minutes

Shared fact sheet: fictional Riverside planting day A youth coalition received permission to hold a planting event in Riverside Square on Saturday. Eighty-five volunteers attended between 10:00 and 14:00. One lane beside the square was closed for 40 minutes while soil was delivered. The city published the closure online two days beforehand; temporary signs were placed on Friday. Police recorded no arrests or injuries. Three drivers submitted complaints about delays. The event created six raised planters; maintenance responsibility is shared by the coalition and parks department.

Intent can be difficult to prove. For youth-work conversations, focus first on evidence, impact and safe next steps rather than accusing a person of deliberate deception. Misinformation and disinformation can both cause harm, but the distinction concerns intent. Misinformation is false or misleading content shared by someone who may believe it is accurate. A useful response is usually to correct gently, provide context, and make accurate information easy to understand and share. Disinformation is created or spread deliberately to deceive. In that case, avoid unnecessarily amplifying it, document relevant patterns, and use trusted reporting procedures. Intent is often difficult to prove from a single post. In youth work, begin with the evidence, the impact, and safe next steps instead of immediately accusing someone of deliberate deception.

Avoid overconfidence in visual detection. Emphasize source tracing, corroboration and context. For sensitive cases involving a young person, preserve evidence and follow safeguarding procedures. Synthetic media changes what can be accepted as visual or audio evidence. AI can generate realistic images of events that never occurred, imitate a familiar voice, alter a person’s face or speech, and produce apparently human online personalities. We should not assume that we can reliably detect these materials because something looks strange. Visual mistakes may disappear as technology improves, while authentic media can also contain unusual details. Stronger verification comes from tracing the source, finding the original context, and looking for independent corroboration. In sensitive cases, preserve the material and follow your organization’s safeguarding and reporting procedures.

Purpose: Practice calibrated uncertainty Time: 15 min. Facilitation: Read the steps aloud. Keep groups focused on observable features rather than guessing motives. Debrief by asking what changed their interpretation and how they could use the activity with young people. You will examine several examples and vote privately on whether each is real, edited, generated, or uncertain. For every judgment, write down the clue you relied on and your level of confidence. After the provenance is revealed, rate your confidence again. Notice whether you were correct for the right reason, correct by chance, or confidently mistaken. The exercise is about calibrated uncertainty, not identifying who has the best eye. “I do not know yet” can be a strong critical-thinking response. Discuss why confidence and accuracy sometimes diverge and what evidence would be needed to reach a firmer conclusion.

Real, edited, generated—or hybrid? Purpose Practise calibrated uncertainty and verify provenance rather than relying on visual intuition. | Suggested time: 15–25 minutes Important In this activity, “real” means a camera-origin image presented without a meaning-changing edit. Routine resizing is ignored. “Edited” means an alteration changes context or interpretation. “Generated” means the depicted scene/person was produced by AI.

Demonstrate with a benign viral claim. Model thinking aloud: what is known, unknown and inferred. Mention reverse image search and source checks without promising that automated AI detectors are definitive. Fact-checking is a short sequence rather than a single magical tool. First, pause and notice whether urgency or emotion is pushing you to repost. Second, trace the claim to the earliest available source and recover its original context. Third, compare it with independent, credible reporting or primary evidence. Fourth, inspect details such as the image, date, author, location, and quoted data. Reverse-image searching may help, but automated AI detectors are not definitive. Finally, decide what action is justified: share, correct, report, archive, or leave the content alone. Unresolved claims should remain unresolved rather than being forced into certainty.

Purpose: Move from suspicion to a justified conclusion Time: 25 min. Facilitation: Read the steps aloud. Keep groups focused on observable features rather than guessing motives. Debrief by asking what changed their interpretation and how they could use the activity with young people. Each team receives a suspicious post or viral claim. Begin by separating the factual claim from commentary, emotion, and speculation. Trace the claim to its original source where possible, and compare it with credible independent information or primary evidence. Record your search so that another person could follow the same trail. Report the result as verified, false, misleading, or unresolved, and explain why. Showing the evidence trail is more important than simply announcing an answer. During the debrief, identify which step was hardest and what additional evidence would be required if the conclusion remains uncertain.

Reflection prompt: Why might a false or misleading story travel quickly? Gather factors from participants—novelty, identity, emotion, simplicity, repetition and social reward—without presenting an unsupported single-cause claim. Strong emotion is a reason to pause, but it is not proof that information is false. Confirmation bias makes information that supports our existing beliefs easier to notice and accept. Fear and uncertainty can make worst-case explanations feel especially plausible. Online environments may also reward fast, confident expressions of outrage more than cautious investigation. Ask yourself: What exactly am I feeling? Which part of the message produced that reaction? What evidence supports the claim independently of my reaction? The goal is not to suppress emotion. Emotion can alert us to values and risks, but it should not substitute for evidence.

Invite examples from youth media. Discuss power: Can a young person freely say no? Is consent specific and ongoing? What happens after content leaves the original channel? Responsible content protects human dignity while still pursuing impact. Consent means people understand how their image, story, or personal data will be used; it should be specific, informed, and ongoing. Privacy means minimizing details that could expose someone to harm. Representation means avoiding the reduction of people to stereotypes, trauma, or decoration. Inclusion requires accessible language, formats, and opportunities to participate. Accountability means supporting claims and correcting mistakes visibly. Before publishing youth-related content, ask whether a young person can freely say no, what will happen if the material travels beyond its intended audience, and whether consent can realistically be withdrawn.

Ask participants to name an inspiring youth-led example from their own context. Focus on what made the process responsible, not only whether the output looked polished. Positive media participation gives young people agency, not simply visibility. Empowerment means young people help define the story and make meaningful decisions about its production and use. Education connects evidence and lived experience in an accessible way. Connection builds understanding across groups without pretending differences do not exist. Amplification increases reach while protecting ownership, consent, and context. Invite participants to share examples of responsible youth-led media from their communities. Focus on the process as well as the finished product: Who made the decisions? Whose voice was heard? Who controlled publication? How were participants protected?

Purpose: Design a message that earns attention ethically Time: 30 min. Facilitation: Read the steps aloud. Keep groups focused on observable features rather than guessing motives. Debrief by asking what changed their interpretation and how they could use the activity with young people. In groups, design a small campaign about mental health, inclusion, anti-bullying, or sustainability. Choose an appropriate format, such as a short video, carousel, meme, or storyboard. Define the intended audience, purpose, and desired action before choosing visual effects. Use at least three modes—such as words, images, sound, color, pacing, or interaction—deliberately. Be able to explain what each mode contributes. Before presenting, complete a consent, bias, privacy, and inclusion check. The challenge is to earn attention without misleading, humiliating, frightening, or exploiting the people the campaign is intended to support.

Run a gallery walk or small-group critique. Require feedback to identify an observable feature and a concrete revision. Creators choose what to change and briefly explain why. Peer feedback helps us test a message before publication. Begin with clarity: What message and requested action do you understand? Next, describe the emotional effect and whether it feels proportionate. Identify claims that require evidence or additional context. Check whether dignity, privacy, and consent are protected. Finally, consider who can access, understand, and feel represented by the content. Use the structure, “I notice… I wonder… A safer or clearer option might be…” Feedback should refer to an observable feature and suggest a concrete revision. The creators decide what to change and briefly explain their reasoning.