Although the literature often uses umbrella terms such as ‘young people’ or ‘youth’—often encompassing ages 15 to 35 (e.g., Duan, 2023; Ford et al., 2024; Juvalta et al., 2023)—this study focuses specifically on ‘young adults’ aged 18–25 due to their distinct cognitive and emotional characteristics (Ecker et al., 2022). This age marks a transitional phase involving identity formation, abstract reasoning, emotional reasoning, emotional regulation, and critical thinking (Schwartz et al., 2015; Simpson, 2018). As part of Generation Z, they are considered digital migrants shaped by technologies and frequently characterised as open-minded and homogeneous (Dimock, 2019; Insider Intelligence, 2023; Hassoun et al., 2023; Parker and Igielnik, 2020). Upon gaining legal adulthood, they can access age-restricted content and be responsible for criminal behaviours (Office for National Statistics, 2019), yet they remain susceptible to online misinformation due to ongoing cognitive development (Enoch et al., 2023; Quintas-Froufe et al., 2024).

Young adulthood is also a sensitive developmental period marked by enhanced vulnerability to environmental stressors (Biddle et al., 1990; Bleidorn et al., 2014). According to the UK’s ‘State of the nation 2022’ report, vulnerable young adults are often targets of coercion and exploitation (Department for Education, 2023), yet they may not recognise abuse or report it, engaging in high-risk behaviours (Arora et al., 2015; Bailey, 2017). Young et al. (2023) found that US college students perceive online harassment as pervasive but exhibit limited engagement in addressing it, often normalising hatred and promoting unjust social comparisons.

Despite awareness of misinformation and fake news, young adults often underestimate its impact. Gurgun et al. (2024) conducted a survey among 247 UK Facebook users and found they are less likely than older adults to challenge misinformation on the platform. A self-reporting survey during the COVID-19 identified young adults (aged 18–35) frequently spread misinformation, often without verifying content (Balakrishnan, 2024). Contributing factors include lack of media literacy education, time constraints, limited interest, and overconfidence in their critical skills (De Vicente Domínguez et al., 2021; Livingstone, 2004; Livingstone et al., 2019; Manca et al., 2021; Xiao et al., 2021).

The critical role of social media platforms in young adults’ communications has also been frequently highlighted in studies with this demographic during COVID-19 (Diepeveen and Pinet, 2022; Ford et al., 2024, 2025; List et al., 2015). Despite heavy use of these platforms, young adults regarded them as some of the least reliable sources of information (Diepeveen and Pinet, 2022). In examining the effectiveness of vaccine messaging from Canadian government Instagram accounts targeting 18–29 year olds for the COVID-19 vaccine, Ford et al. (2024) highlight that these efforts failed to meet the needs of this demographic, particularly in building trust and addressing online misinformation.

However, more research is needed to clearly define the contexts in which young adults encounter online misinformation and fake news, and how they make sense of it, beyond just understanding their general perceptions and attitudes. This will help identify specific risks in concrete scenarios and inform the development of more targeted interventions.

Yu et al. (2023) conducted a systematic review of 64 empirical studies to understand strategies for reducing misinformation sharing and promoting correction sharing in today’s decentralised informational landscape. Their findings highlight several social media platform features that contribute to the recirculation of misinformation. These include credibility indicators such as user and expert reputation ratings (Kim et al., 2019); virality metrics like the number of retweets (Lee and Oh, 2017), likes and replies (Kim, 2018); the influence of power users (Shao et al., 2018); profit-driven algorithms (Braun and Eklund, 2019); and the employment of bots (Shao et al., 2018). The sequence of exposure to rumours and rebuttals also affects message perception and acceptance (Pal et al., 2020). These findings align with earlier research by Bryanov and Vziatysheva (2021), which explained why individuals fall for fake news and the factors influencing belief in misinformation. For instance, message topics and bandwagon heuristics, represented by the numbers of likes, can significantly enhance the perceived credibility (Luo et al., 2022). Yu et al. (2023) also found that perceived message importance mediates misinformation and correction-sharing behaviours (Oh and Lee, 2019), aligning with prior work on the role of innate beliefs and ideological predispositions (Pennycook and Rand, 2020). However, more research is needed to explore how young adults assess online information in everyday contexts and what actions they take based on trust or mistrust.

Existing studies in the social sciences and HCI have examined practical approaches to help a broader demographic identify and prevent the spread of misinformation and fake news online (e.g., Hassoun et al., 2023; Nazari et al., 2022; Staender and Humprecht, 2023; Veeriah, 2021). Various measures for fact-checking have been compared, designed, and evaluated, including the use of traditional news media, fact-checking organisations, professional journalists, as well as emerging approaches such as Artificial Intelligence (AI) and crowdsourcing (e.g., Almeida et al., 2023; Haque et al., 2020; Li and Chang, 2023; Liu et al., 2023; Santos and Pereira, 2024; Shusas, 2024). Sehat et al. (2024) proposed a priority framework to assist fact-checkers in making strategic decisions to evaluate and address the harms caused by misinformation. Other measures include interventions by trusted individuals, such as family members, who challenge the misinformed person offline—a strategy highlighted as effective by Scott et al. (2023). Additionally, games have been developed to raise awareness about the tactics employed by misinformation creators (e.g., Humphrey, 2023; Karlova, 2018; Maertens et al., 2021; Roozenbeek and Van Der Linden, 2019). Warning labels designed to discourage users from engaging with misinformation or assist in correcting it, along with preventive measures integrated into messaging platforms to limit content sharing, have also been explored (e.g., Broniatowski et al., 2023; Nassetta and Gross, 2020). However, some interventions, such as setting warning flags, may only be effective for a short period (Broniatowski et al., 2023; Scott et al., 2023).

Whilst researchers have started to explore specific vulnerable groups, such as older adults and individuals who have previously shared low-quality news (e.g., Brashier, 2024; Moore and Hancock, 2022; Peng et al., 2024), there remains a lack of evidence addressing the unique implications for young adults. Teenagers are growing up in an era where information on social media saturates their lives, and influencers significantly shape their options (e.g., Peter and Muth, 2023). We consider our work both timely and critical in raising awareness of these experiences. Furthermore, we offer considerations for stakeholdesrs in global societies who are involved in the health and growth of younger generations.

We define ‘agency’ as the subjectivity of individuals in holding motivational attitudes and taking initiative aligned with those attitudes (Hainz et al., 2016). In our study, we focus on young adults’ proactive attitudes and their capability to act against the spread of misinformation and fake news they encounter online. We emphasise the importance of their ability to shape outcomes rather than merely exchange information, whilst also evaluating such behaviours with regards to their own beliefs and attitudes (Hainz et al., 2016; Lukoff et al., 2021). Bennett et al. (2023) categorised human agency into four broad aspects: (1) the level and directness of casual involvement or value alignment (e.g., Birk et al., 2016; Lazar et al., 2017; Mentis et al., 2019; Valencia et al., 2020); (2) the sense of agency in triggering events (e.g., Kasahara et al., 2019); (3) the timescale of agency manifests, from momentary to lifelong (e.g., Cornelio Martinez et al., 2017; Foley et al., 2019); and (4) individual independence versus social interdependence (e.g., Garg, 2022; Güldenpfennig et al., 2019). Despite these categorisations, researchers often fail to articulate or examine the coordination and trade-offs between these dimensions (Bennett et al., 2023). Jicol et al. (2021), for instance, explored emotion and agency in Virtual Reality settings, showing that agency moderates the emotion’s effect on presence.

In the social sciences, particularly in Social Media Studies, agency has been examined through diverse lenses. Increasing attention is directed towards agency involving machines and algorithms in mainstream media and journalism (e.g., Linden, 2017; Rydenfelt, 2022; Verwiebe et al., 2024). Researchers have argued to maintain human oversight and control by ensuring individuals are informed about and capable of monitoring the behaviours of automated algorithmic agents, which should not independently make decisions (Linden, 2017; Rydenfelt, 2022). Complementing this perspective, Clark et al. (2015) explored the potential for agency in generation of young people through exchange storytelling. Their work highlighted how digitally enabled narrative exchanges on digital and social media platforms can foster a deeper awareness of the complex contexts that young people connect with.

Existing research across different fields is establishing connections. For instance, Bennett et al.’s (2023) generation of the interdependence within social contexts aligns with Clark et al.’s (2015) proposition of encouraging such awareness in education to generate agency among young people. However, a notable gap remains in further identifying other dimensions of agency that contribute to positive outcomes, such as enhancing individuals’ sense of self and wellbeing (Bennett et al., 2023; Lukoff et al., 2021). We recognise an opportunity through our recent work to contribute new, empirically grounded implications for how to promote agency in ways that enable young adults to protect themselves whilst mitigating the harms and risks associated with navigating information online, especially when encountering online misinformation and experiencing forms of online harm.

The sample comprised: nine external stakeholders with interdisciplinary backgrounds across private sector, NHS, Higher Education, the Voluntary, Community and Social Enterprise (VCSE) sector, and the public sector; two young adult representatives; and 13 internal project researchers. We held the workshop in a hybrid format (i.e., three offline groups and one online group), with a meeting room organised into roundtables. In accordance with the informed consent and workshop protocol, we assigned participant IDs to each participant along with a brief description of their role and expertise (see Table 1).

We engaged in discussions in response to three prompts across three round-table sessions. Each group appointed one researcher participant, to act as a coordinator and responsible for taking notes. At the end of each session, one participant from each group was selected to summarise the discussion outcomes on behalf of the group members. Following each presentation, all participants were encouraged to share their reflections.

After this workshop, we transcribed the audio recordings and collected hand-written notes captured at each table. We employed a Deductive Reflexive Thematic Analysis (RTA) method (Boyatzis, 1998; Braun and Clarke, 2012, 2019; Roberts et al., 2019), to capture, structure and interpret the data in relation to our predefined research questions, drawn from participants’ rich discussions and reflections (Braun and Clarke, 2006; Naeem et al., 2023). Audio recordings were transcribed by a trusted third-party company.

The first author of this paper began by thoroughly reviewing the transcripts to become familiarised with the data. The transcripts were then imported into NVivo for coding. Based on the existing research questions, the first author developed a set of preliminary codes and applied them to relevant segments of data. These codes were then subsequently organised and connected, facilitating the identification of recurring themes and patterns that aligned with the study’s focus.

The initial themes and corresponding coded excerpts were exported into an Excel file for further review and refinement. In line with RTA, the authors collectively reviewed the clustered themes during weekly meetings, fostering critical reflection and ensuring that the themes accurately represented the research questions.

In the second workshop (W2), the participants were assigned with participant IDs W2P1-W2P7. Our co-design method appropriated Ideation Card Design Method (Golembewski and Selby, 2010; Hope and Mulhall, 2024) and Storyboarding Techniques (e.g., Greenberg et al., 2012). Our objective was to facilitate the participants to collaboratively create (co-create) specific user scenarios of interacting with online technologies that demonstrate and depict their experiences mediated by online media consumption. Our appropriation of the Ideation Card method (Golembewski and Selby, 2010) utilised visual storyboarding (e.g., akin to comic strips) to depict the scenarios where individual (fictional) characters may encounter (or have encountered) online misinformation and fake news, and where interactions with online technologies can take place between particular people; these scenarios could be fictional but inspired by or grounded in participants’ personal experiences or the experiences of their peers. The rationale for using this method was to focus discussion on user experiences with technology (ibid). The broader rationale for the co-creative setting and use of storyboarded, character-driven scenarios was to prompt dialogue and reflection by young adults on specific contexts of interaction, to raise considerations for technology designers (Durrant et al., 2018) (see Table 2).

We now report on the storyboard contents and illustrations co-created and narrated by our participants in the co-design workshop.

Our findings highlight that young adults’ emotions and aspirations are often exploited in various contexts, plus their frustrations in seeking healthcare support, anxieties about fitting into social circles, efforts to secure employment, and desires for quick financial gains. In these situations, their eagerness to engage and succeed can increase their vulnerability to harm, particularly through the dissemination of misinformation by predatory ‘bad actors’.

In discussing the spread of healthcare misinformation, W1P8 (NHS) described a scenario in which young adults ‘get progressively frustrated by being passed from service provider to service provider’, after ‘following the traditional route to access a health service for a particular health problem’. The frustration and eagerness may drive them to seek out alternative support, which can result in serious harms. This NHS worker provided an example of ‘people not getting a particular medication that they want prescribed, they may go and source it from alternatively online through illegal sources’.

We find that such urges to seek healthcare support can be taken advantage of by certain ‘bad actors’. The public media corporation representative (W1P7) highlighted the allure of ‘capitalism and profit-making’, which draws in certain ‘hidden game changers’—some pharmaceutical companies involved in collecting and monetising personal health data. These companies crafted targeted misinformation online, influencing and disrupting young adults’ online efforts to seek reliable information and support.

Like the protagonist Gavin in the hair growth advertisement scenario, participants from the co-design workshop (W2) noted that young adults can be easily affected by social anxieties, and by the efforts made to fit in social circles. Hence, they tend to seek simple, ‘One for all’ (W2P3) solutions to their problems, frequently neglecting to invest time in scrutinising product descriptions or verifying the source of advertisements they encounter.

Social media platforms have arguably capitalised on the concerns and peer pressures experienced by young adults, by offering faster and more convenient pathways to online purchasing. As W2P4 noted, ‘If you see it on a video or in a post, you can just literally click on a link, and it will take you to the shop that’s within the social media app’. This can be exacerbated by ‘celebrities and influencers’ whose revenue depends on attracting audiences, promoting, and selling products (W1P12, computer science). The selected publication channels, identified as ‘social mediums’, are often ‘less inclined to moderate the content and advertisements’. W1P5 (postgraduate) added: ‘Sometimes they are less inclined to moderate what advertisements they are actually selling through their promotions.’ Arguably, such commercialised online environments further influence young adults’ against applying logical judgment and critical thinking when navigating digital spaces.

In the ‘Employment seeking advertisement scam’ scenario, W2P5 highlighted that fake job scams would normally ‘misplace letters’ of renowned company names in order to target and mislead young adults who do not identify carefully when looking for a job, prompting them to think that the opportunity is with the renowned company. Refinement on the details such as reputation and history of fake companies would make it more difficult for young adults to distinguish information veracity, because ‘people thought it was reputable, so they didn’t have any doubts’ (W2P5). Young adults, in their desire to seize and secure job opportunities, may share personal information, such as identity and bank details, through the convenience of auto-complete browser functions. They seem to be susceptible to this when they want to secure a job position, and do not take the time to verify a company’s legitimacy, potentially leading to financial loss, data breaches, and an increased risk of future ‘scam calls’.

W1P3 (Digital Identity Representative) cited the example of young adults being harmed by ‘money muling’¹ advertisements, which can attract individuals with the promise of immediate financial gain whilst obscuring the potential long-term subsequent consequences for themselves and others:

Hence, W2P3 and W2P4 proposed that designers should consider how to stop young adults from making rushed decisions when what they ‘normally look for efficiency’. Technology applications should be designed to prompt their users to make time to ‘pause’, to reflect on their own judgment, ‘rather than being persuaded’ (into a rushed decision).

W2P4 also observed that, everyone is susceptible to ‘emotional moments’ and may experience ‘certain degree of disabilities’ at times. Young adults in particular find themselves experiencing heightened emotional states, ‘such as feeling isolated at home or about stressed job security and happiness’ (W2P4). In such states, individuals are more likely to be harmed by misinformation and fake news, as they will ‘feel difficult to identify them’ (W2P4, W2P7). Consequently, they may be easily targeted or manipulated, as misleading information can offer ‘a false sense of control’ (W2P7). The emergence of AI algorithms raises new concerns, particularly around how new media technologies like chatbots may exploit young adults. These AI-driven technologies can project positivity, persuasiveness, and a ‘keenness to help’ (W2P1) by actively listening and providing support during young adults’ expressed emotional needs. W2P3 commented that unlike ‘a friend or a family member who may really want to help but don’t have the information’, such technologies can easily ‘draw people in’ with tailored responses.

In summary, this section reported how young adults’ expressed emotions such as frustration, eagerness and anxiety can be exploited by targeted misinformation, especially during times of mental stress, physical discomfort, peer pressure, or financial instability. Participants noted that such potential harm from misinformation encounters could extend to anyone experiencing heightened emotional states at certain points in life. In these contexts, misinformation targeting is shaped not only by content but also by underlying social dynamics.

Across both workshops, participants agreed that young adults’ trust in online information is highly susceptible to influence. Several contexts were described to illustrate the factors affecting trust, such as the perceived reliability of information, the influence of endorsements from trusted family members, the reactions of the public, and long-term influences by closed online communities dominated by singular and uniform opinions or voices.

In the co-design workshop (W2), young adults shared their curiosity about how peers assess the trustworthiness of health information. They observed that chatbot-generated responses might more readily earn users’ trust because of the chatbot’s anthropomorphic language use, making them more approachable and reliable. When searching for similar questions on search engine, participants noted that the results are often presented in an exaggerated manner. The outcomes can be fear-inducing and overwhelming.

They also highlighted the difficulty of evaluating information provided by the ChatGPT chatbot. In particular, the ‘one-time’ (W2P2) and real-time nature of the communication makes it challenging for users to revisit and re-evaluate the previous responses provided as ‘there’s no coming back to questioning it’ (W2P4). Hence, it is also difficult to lose the trust of human users once it proves its efficiency.

Trusted close family members, such as the brother in the ‘hair growth advertisement on Instagram’ scenario, can easily influence young adults’ decisions to trust when faced with misleading advertisements, ‘even though the (protagonist’s) brother clearly doesn’t have any knowledge around hair growth’ (W2P3). To young adults, such trusted people can ‘act as a proxy who is able to, hopefully, understand the world better than them’, and let the demographic ‘put down their guard’.

Young adults’ trust to official information sources can be easily swayed, particularly when mistrust in health experts spreads rapidly through social ‘contagion’ (W1P14, law). For example, W1P14 spoke about ‘vaccine hesitancy’ during the Covid-19 pandemic as an example, describing how circulating doubts and mistrust towards health authorities led to widespread reluctance or refusal to get vaccinated. Additionally, public signals reflected by ‘likes’ or ‘votes’ on social media platforms (e.g., Instagram) can influence young adults’ trust in online content. As W2P5 commented, ‘some young people see the number of likes as a sort of, vote of confidence’.

Another high-risk context for young adults’ exposure to online misinformation is within echo chambers. In these spaces, young adults’ trust in and judgment of others can become desensitised, often leading to the normalisation of inappropriate behaviours. W1P12 (computer science) noted that individuals in echo chambers ‘only got a partial awareness’ and ‘tend to only look at news or only hear the views of people that agree with them’. Some young adults who ‘believe everything is fine’ will gradually stop questioning information validity.

In these closed communities, AI-generated ‘personalised fake accounts’ driven by ‘political or societal campaigns’ can effectively propagate misinformation to align with campaign goals. A significant challenge arises in tracking and preserving evidence within these real-time communication channels, making it challenging to counter misinformation. For example, W1P14 speaks about such an occurrence in the context of online gaming communities:

Young adults’ criteria for trust are found to be varied and often unpredictable. They may endorse trust in AI tools, close family members without specific expertise, or be swayed by public opinion, sometimes even challenging authorised and professional insights. Echo chambers can desensitise young adults to differing perspectives, making it harder to identify misinformation and engage with diverse viewpoints.

Participants in the stakeholder engagement workshop highlighted several marginalised groups that are often overlooked by mainstream society and are particularly vulnerable to online misinformation. For instance, W1P14 (law) mentioned ‘online gaming or virtual reality gaming’ communities often marginalise or trivialise certain groups, particularly ‘women or girls, and individuals from diverse racial backgrounds’. Such communities may perpetuate stereotypes or dismiss the experiences of these groups, reinforcing the exclusion of their voices. Hence, participants called for more inclusive research engagements with young adults whose voices are silent or silenced within online communities, to illustrate both idiographic journeys and experiences of complex online harms:

W1P1 (Police) expanded the discussion of marginalised audiences to include not only young adults but also children. He highlighted concerns around ‘an organised criminality harvesting the information of children, manipulating and engaging with them’. He raised an example of Kinder eggs to illustrate these tactics:

By identifying contact details of children who show interest through specific keyword searches on search engines, organised networks can effectively build a customer base. This enables stakeholders to ‘target their campaigns to children’ (W1P3) with particular vulnerabilities, minimising the need for direct targeting efforts.

In addition to specific gender, age, and racial groups, W1P3 (Digital Identity) highlighted the cascading chain effects of familial and social challenges faced by some young adults, which heighten their susceptibility to problematic situations. She noted that interconnected incidents are like ‘chinks in the armour’ that can exacerbate the fragility to misinformation and other online risks, such as being ‘neurodiverse’, experiencing ‘a parental break up’, or having ‘a parent or sibling going into prison’. Personal and social struggles often compound and limit their resilience and access to reliable support. She explained the reason:

Malicious actors can exploit marginalised demographics with specific vulnerabilities by using AI algorithms to ‘ascertain characteristics such as gender, age, cognitive or physical disabilities’ (W1P3). This enables them to target misinformation more precisely at the selected audiences, increasing impact of misleading or harmful content.

Another form of vulnerability arises from the challenges of understanding complex medical, legal, or commercial terminologies online. Participants (W1P8 NHS, W1P14 law, W1P20 media) in the stakeholder engagement workshop (W1) emphasised that such language barriers prevent young adults from fully ‘understanding a situation’ and often leaving them ‘trying to make sense of in a given moment or situation’. They highlighted the example of young adults seeking healthcare support online. Given that online spaces provide healthcare information from diverse sources worldwide, often in different languages and cultural frameworks, the information exists in a legally ambiguous juridical territory. The space ‘is no longer governed by the best advice of a UK health provider’ (W1P14), making it challenging to safeguard young adults from unverified or potentially misleading healthcare sources, which may impact their wellbeing:

Young adults are vulnerable to spread of misinformation when online bystanders amplify the situation, particularly outside school spaces. W1P1 (Police) expressed concern about how content regarding conflicts between young adults can escalate and spread rapidly online. Online bystanders, such as ‘parents or just people who are on the periphery’, or those ‘just out of meanness who what to stir the pot’, can ‘add fuel to’ the fire, exacerbating the situation without considering the consequences. As W1P3 (Digital Identity) noted, people may not realise that what they are sharing is ‘fuelling, supporting and committing harmful conduct’.

In sum: across both workshops, participants identified several types of vulnerabilities that increase young adults’ susceptibility to misinformation. These include marginalised demographics, challenges in understanding complex legal, medical, or commercial terms across online boarders, and the potential harm caused by online bystanders outside of school-supervised environments. Such factors contribute to the potential risks young adults face online, as each can increase their exposure to and impact from misinformation.

To address the prevalence of misinformation, both W1P7 (Public Media) and W1P3 (Digital Identity) stressed the importance of equipping young people with personalised techniques and countermeasures to effectively ‘debunk and identify’ the harmful practices and strategies employed by bad actors. They highlighted the necessity to incorporate these resources into school curricula. Promoting media literacy, they argue, offers ‘a better way of trying to understand the spread of misinformation and who is impacted’.

W1P14 and W1P20 (Law and Media) highlight the importance of considering the ‘rationality and contextual nature’—the contextual significance of ‘agency’. They argue that solely increasing young adults’ agency is insufficient, as it can wrongly validate individuals’ perceived expertise on complex issues in which they personally lack knowledge, thus heightening their susceptibility to the influence of misinformation. Hence, the extent of agency granted to young adults in navigating online information and using digital devices requires reconsideration within societal disclosure:

Nevertheless, participants discussed how encouraging individuals to develop self-responsibility by promoting their agency is not the solution to mitigate against harms. In addition to endorsing individual agency, W1P20 (Media) also emphasised the importance of ‘collective agency’ (Hainz et al., 2016), which can foster group-based critical thinking about interventions and responses that can enable both informal and formal social controls. One example is the inclusion of oversight by mainstream journalists:

The discussion of traditional gatekeepers in media accuracy such as news reporters and journalists prompted participants to explore evolving definitions of ‘guardians’ in the context of social media. W1P5 (postgraduate) noted that some young adults see themselves as guardians of safe online spaces, taking on roles in ‘advocating online safety’ and ‘providing peer support’. They actively monitor the veracity of information and exercise agency over how it is shared.

This theme highlighted participants’ insights on key directions social media platform developers and policymakers should consider reducing young adults’ exposure to and harm from misinformation. Discussions emphasised the importance of empowering young adults with agency to identify and stop the spread of misinformation, whilst also acknowledging challenges that may require additional intervention from mainstream gatekeepers and other stakeholders. In addition, providing more tailored materials to enhance young adults’ media literacy skills in identifying misinformation was seen as critical.

During the first workshop (W1), stakeholders (e.g., W1P12) discussed the serious harm and risks posed by the spread of misinformation within small, enclosed online communities—often referred to as echo chambers or filter bubbles (e.g., Pennycook and Rand, 2020; Kaluža, 2022). Both terms are normally conflated in existing work and are frequently driven by social or political campaigns and exacerbated through the curation and manipulation of algorithms. Social media and other new media platforms afford both the representation and provocation of emotions, which can significantly influence the escalation and de-escalation of violence (Morales, 2024). The amplification of emotional responses and the resulting inappropriate behaviours, either through heated discussions or fabricated contents tailored specifically for the echo chamber spaces, can fuel conflicts by reinforcing polarised narratives.

An abundance of literature has explored and analysed how misinformation is created and disseminated to desensitise young adults within echo chambers. For instance, Diaz Ruiz and Nilsson (2023) highlighted that the malicious actors would strategically ‘seed’ the deceptions within these spaces by masquerading misinformation as credible information. Once introduced, members within the echo chambers are enlisted to cocreate and propagate these contentious narratives, further disseminating the misinformation. As emphasised by W1P12, though initially seeking peers to share common interests, young adults may narrow their worldview to align solely with the perspectives within these spaces, leading to the subjective rejection of diverse voices and the gradual loss of critical thinking. However, there is a notable scarcity of empirical evidence involving users within echo chambers. Existing work, such as Söderberg's (2016) interviews with 13 users of anti-immigration internet media, explored why individuals consume such content and whether they are influenced by echo chambers. Despite this, it remains challenging to track and understand how young adults are gradually influenced within closed, encrypted groups like Telegram (Molla, 2021), where the dynamics of misinformation propagation and ideological reinforcement are often concealed. We propose adopting research methodologies such as data diaries (Vlachokyriakos et al., 2024), which allows for the longitudinal collection of self-reported data, to better track and understand young adults’ behavioural change within such spaces over time. Notably, engaging with this demographic presents challenges, as young adults within echo chambers may be reluctant to participate in research activities or may resist acknowledging the presence of misinformation (Du, 2023, p. 20).

Participants also specifically talked about online gaming as a critical area for future research and engagement. The communication environment in gaming is often characterised by hostile and aggressive expressions, including discrimination, racism, and sexism (Meriläinen and Ruotsalainen, 2024). Due to the online disinhibition effect, young adults may feel less socially restrained, leading to more uninhibited and open expressions (Suler, 2004). Furthermore, the gaming environment can normalise such behaviours, making young adults perceive them as more acceptable (Meriläinen and Ruotsalainen, 2024). Compounding this issue, messages conveyed through voice are particularly difficult to capture as evidence, whilst tone and emotional delivery can more directly provoke responses and reactions (Lawson and Mayer, 2022). Given these dynamics, we believe this is an essential field requiring the involvement of researchers and educators to better understand and address the complexities within this unique context.

In addition, online technologies that enable encrypted communication and anonymity further complicate the situation within echo chambers by allowing individuals to engage in harmful activities without the risk of incrimination (Walther and McCoy, 2021). As W1P14 highlighted, AI-generated personalised fake accounts further contribute to the blend of voices in these enclosed spaces. Such AI deception—where AI systems are trained with malicious intent—can be seen a process of human disempowerment (Park et al., 2024). In these echo chambers, intentionally created to attract vulnerable groups of young adults seeking peer support, voices gradually become contaminated by increasingly radicalised and polarised voices and opinions if they operate without supervision or moderation.

Hence, there is an urgent need to address how misinformation proliferates in these spaces and to evaluate the effectiveness of potential interventions through engagements with young adults and related stakeholders. By conducting empirical research in this area, researchers and related stakeholders can propose more effective policies and design tools that better safeguard young adults from these digital risks.

In this section, we consolidate insights from the workshops into strategic design implications.

Our qualitative approach for both workshops determined a relatively small, non-representative UK study sample and could be complemented by a larger-scale future study with a representative sample. Also, in the co-design workshop (W2), our choice of co-creation methods led to our recruitment of participants who have specialist expertise in Interaction Design. This focus on expert participants introduced a bias, because it engaged young adults with higher levels of media literacy and technology skills. Future research could involve a more diverse sample of young adults.

In terms of extending the qualitative research on this topic and reflecting on methods: whilst the storyboarding techniques drew anecdotally on lived experiences, there is also an opportunity to use different speculative design methods that engage research populations in the real-world contexts of their lives, for example, more performative techniques (Elsden et al., 2017; Hutchinson et al., 2025), or other real-world interventions and observations (Cipolla, 2018). Such methods could invite co-creativity and speculation in participants, whilst also capturing direct experiences of media creation and consumption that could deliver evidence on situated behaviours and sense making.