AI's rapid advancement offers both opportunities and challenges for older adults, who risk being marginalised by technologies they could otherwise benefit from. Age-Friendly AI, Ireland's National AI Literacy Programme for Older Adults will engage 60,000 older adults across Ireland in a conversation about AI and its relevance to their daily lives. The project is split into two phases. Phase 1, currently underway, uses a co-creation approach via a series of interactive, multi-stakeholder workshops involving older adults, organizations that support them and policymakers. Based on the insights gathered from these interactions, a tailored and accessible AI literacy programme will be co-created. In phase 2, the co-created 'train-thetrainer' programme will also be hosted with partner groups in libraries across Ireland to equip community facilitators, librarians and age-friendly organisations with the skills and resources needed to deliver the AI literacy programme to older adults in their local communities. In this position paper, we present our methodology for engaging stakeholders at co-creation workshops and present initial findings from surveys that were administered and discussions that were facilitated at these sessions. Our findings highlight that older adults recognize the potential of AI to support independence, healthcare and social connection while also raising concerns about privacy, trust, bias and the preservation of human interaction. We discuss how these insights will shape the development of an accessible AI literacy programme for older adults.
As artificial intelligence (AI) becomes increasingly embedded in everyday technologies, it is important that older adults are not left behind in the digital transition. While AI offers significant opportunities to support independent living, enhance healthcare, improve family and social connectivity and access to services, older adults often face barriers to understanding and engaging with these technologies. This exclusion risks deepening existing digital inequalities.
Research shows that digital technologies can support the everyday lives of older people and
contribute to a better autonomy, well-being, and quality of life as well as reduce loneliness [1, 2, 3,
2nd Workshop on Education for Artificial Intelligence (edu4AI 2025, https://edu4ai.di.unito.it/), Co-located with ECAI 2025,
the 28th European Conference on Artificial Intelligence which will take place on October 26, 2025 in Bologna, Italy
∗ Corresponding author.
† These authors contributed equally.
1.1.
4, 5]. With the rapid rise of AI technologies and their increasing integration into daily life, AI literacy
has become an important extension of digital literacy. There is a new need for programmes to
promote AI literacy and competency to encourage adoption of AI technologies to enable better
participation in digital society for older adults [
This paper introduces Age-Friendly AI, Ireland’s National AI Literacy Programme for Older Adults, which aims to bridge this gap by grounding AI education in human-centred design and cocreation. The project brings together older adults, age-friendly organisations and policymakers to shape a national conversation about AI from the ground up and to co-create content for an AI literacy programme specifically tailored to older adults’ requirements. The programme will aim to educate older adults in AI concepts with the aim of building confidence in discussing AI and in using AI systems and application with critical awareness. Phase 1 of the project uses a co-creation approach through a series of interactive, multi-stakeholder workshops involving older adults, organizations that support older adults and policymakers. Part 1 of phase 1 focuses on co-creation sessions with older adults to capture lived experiences, hopes, concerns and expectations about AI. In part 2 of phase 1, these insights will form the basis of a series of further workshops that will bring together older adults, community organizations and policymakers. These sessions will create a structured space for dialogue, enabling stakeholders to critically reflect on the findings from part 1 and to identify shared priorities. This paper describes our co-creation methodology and initial findings from co-creation sessions from part 1 of phase 1 of the project, which consisted of co-creation sessions with older adults. We outline how these insights will shape further sessions with community organizations and policy makers and the development of an accessible AI literacy programme for older adults.
Digital inclusion promotes equitable access to technology services, skills and opportunities,
enabling everyone to participate fully in the digital economy and society. However, Age Action, an
Irish NGO, reports that 62% of adults in Ireland over 65 lack basic digital skills, with nearly one-third
not using the internet at all [
To mitigate these gaps, digital literacy programmes tailored to older learners have been
developed, emphasizing collaborative, cooperative, and intergenerational learning which are
strategies found to boost engagement, creativity and social interaction [
Despite the prevalence of digital literacy efforts, fewer initiatives combine AI-specific learning with multi stakeholder co-creation at scale. This is largely because AI, particularly generative AI, has rapidly emerged in recent years, outpacing the development of tailored educational programmes. Moreover, AI has become highly visible in media and public discourse, intensifying awareness and concern across all age groups. This gap is significant because AI technologies increasingly impact everyone, including older adults, who face unique challenges and opportunities in navigating these advances. The Age-Friendly AI project aims to address this need by integrating diverse stakeholders within a collaboratively co-designed, evidence-based AI literacy programme for older adults.
The rest of this paper is organized as follows. In the methods section we outline our co-creation process including participant recruitment, interactive workshops and data collection techniques. In the results section, we present initial findings and stakeholder feedback from co-creation sessions. In the discussion section we describe how we will contextualize data and information from the cocreation phase to develop, deliver and evaluate a national AI literacy programme for older adults.
The Age-Friendly AI project is part of a larger EPE (Education and Public Engagement) programme within the Research Ireland ADAPT Centre for AI-Driven Digital Content Technology [17]. In 2021, ADAPT launched #DiscussAI [18], a national conversation designed to build AI literacy and empower people across Ireland to critically engage with AI-driven systems.
The #DiscussAI national conversation was recognized in 2025 by the European Commission as a best practice in science engagement [19]. The core aims of #DiscussAI include: • Educate the public about AI and develop core media literacy skills in the context of AI • Enable the public to contribute to a national dialogue on social impacts of AI • Ensure that diverse voices and experiences shape the future of AI research and policy in Ireland • Build shared understanding of AI among citizens, researchers, educators, and policymakers The methodological framework for the Age-Friendly AI project is grounded in the practical insights and lessons derived from the #DiscussAI programme. The co-creation activities involving older adult participants follow a carefully designed workflow, organized in collaboration with our established networks of partner community services and strategically held in local community settings across Ireland:
Our participant recruitment strategy was designed to achieve broad reach and inclusivity across Ireland. Participants were recruited through a purposive sampling approach. This involved leveraging our established networks of partner community services. These partnerships ensured access to a wide demographic range across Ireland ensuring representation in various age ranges (65+, 75+, 85+), digital literacy levels, socioeconomic backgrounds, and geographic distribution (urban/rural).
2. Choosing a venue that is easy to reach by public and private transport and equipped with accessible facilities to ensure comfort and inclusion for all older adult participants.
We collaborated with local community organizations to host workshops in familiar, trusted venues that, where possible, participants already knew and felt comfortable visiting.
A pre-event survey is administered to systematically gather baseline insights into participants’ prior knowledge of, confidence in engaging with and talking about AI technologies, products and services. This provides a baseline for subsequently measuring the impact of the workshop. 4. An introductory presentation that includes: • An overview of AI: Demystifying what AI is, its fundamental principles, and its rapid evolution. • Real-world examples: Providing concrete, relatable examples of where AI is already embedded in everyday products and services that participants might use or encounter, such as Google Maps for navigation, predictive texting on smartphones, and the AI algorithms behind banking card payments and fraud detection. • More complex AI concepts: Introducing more nuanced and challenging AI concepts, including the phenomenon of AI bias (how biases in training data can lead to unfair outcomes), AI hallucinations (where AI generates convincing but false information), and the significant environmental impact of AI training and processing (e.g., energy consumption of large language models). 5.
Following the introductory presentation, participants move to smaller discussion groups, with five or six individuals per table. Each table is supported by a dedicated facilitator, who has undergone a 2-hour facilitation skills training focused on active listening, managing group dynamics, ensuring equitable participation and maintaining neutrality. Their role extends to probing for clarification, gently steering conversations back to the core themes, and ensuring a respectful and inclusive environment. A designated scribe at each table is responsible for capturing key discussion points and areas of consensus or divergence, providing a rich qualitative data for subsequent analysis.
As part of the session (before and after the structured components), participants are provided with a hands-on opportunity to interact directly with a number of AI demonstrations. These include: • • •
The ’Art or AI?’ exhibit developed as part of the #DiscussAI programme, which challenges participants to distinguish between images of human-made art and AIgenerated art [20]. This exhibit enables participants to develop a foundational understanding of the capabilities and limitations of generative AI, to critically evaluate AI outputs and engage meaningfully in broader societal conversations about the ethical development and responsible integration of AI technology. The open source ’Art or AI?’ exhibit and quiz can be accessed digitally at: https://bit.ly/ArtorAI The ‘AI Mask’ exhibit, an animatronic ’talking’ mask which is a physical embodied conversational agent [21]. The mask provides an accessible, tangible and embedded interface to embody AI features and technologies including Large Language Models (LLMs), generative AI, machine learning, natural language processing, and computer vision systems. This exhibit is used to showcase and provoke conversations about AI technologies as well as broader discussions on AI’s role in creative processes, its ethical implications, and its effect on society in general.
The ‘Everyday AI’ demonstration is an interactive showcase of widely available generative AI tools such as ChatGPT and DALL·E. This exhibit highlights how conversational AI and image generation systems can be used in accessible and creative ways, from answering questions to producing visual content from simple text prompts. Visitors are invited to engage directly with these tools to better understand how they operate, their capabilities and limitations and the kinds of opportunities and challenges they present. The demonstration encourages critical reflection on how generative AI is becoming embedded in daily life, raising questions around trust, bias, creativity and responsible AI use. 7. Post-event survey
The co-creation workshops conclude with a post-event survey. This survey mirrors the pre-event survey in structure but is aimed at gathering participants’ quantitative and qualitative insights into their experience with the workshop, feedback on the materials and exhibits and any experienced changes in their knowledge, confidence and perceptions of AI following the interactive sessions.
Since January 2025, the project has engaged a total of 252 older adults who have participated in cocreation workshops across five regional locations across Ireland. The age and gender distribution of the participants are shown in Figure 1 while Figure 2 demonstrates educational background and selfreported digital skills levels. Figure 1 shows that the largest proportion were aged 70–79 years, followed by those in the 60–69 group, with a smaller representation from the 50–59 and 80–89 age groups. Figure 1 also reveals a clear gender imbalance, with females making up over 80% of the population compared to a small proportion of males. This is in line with other studies that have found it is common for women to participate in community-based studies at higher rates than men, a pattern observed in social science and other participatory research settings (e.g. [22, 23]).
Figure 2 shows the distribution of education background and digital skills levels among the surveyed participants. The majority (around 70%) have a third-level education, while smaller proportions have secondary education (about 21%), vocational training (roughly 4%) or other forms of education (4%). In terms of digital skills, most participants report good skills (49%), followed by a moderate number who rate their skills as neither good nor poor (32%). Smaller groups report very good digital skills (about 8%), poor skills (9%) or no experience (4%). A comparative analysis of pre- and post-event survey responses for a sample of 81 participants revealed significant improvements in participants’ reported perception of their understanding of AI and confidence in sharing their views about AI. The comparison is presented in Figures 3 and 4.
From Figure 3, the pre- and post-survey results reveal an increase in participant’s baseline understanding of how AI works. Before the event, most participants rated their knowledge as “Very Poor” or “Poor” (43.8%), with only 10.1% feeling “Good” or “Very Good.” Following the event, the majority (88.6%) rated their understanding as “Good” or “Very Good,” while only 3.8% remained at the “Very Poor” or “Poor” level.
The results from Figure 4 indicate an increase in participants’ confidence to share their views on AI following the event. Before the event, 49% of participants felt “Comfortable” or “Very comfortable,” while 35.6% were neutral or uncomfortable. After the event, this proportion rose to 80.7%, with only 2.6% feeling “Very uncomfortable” or “Uncomfortable.”
From the facilitated discussions, participants recognized both challenges and opportunities in the adoption of AI technologies. Key concerns include privacy and security, with worries about data misuse, scams, and lack of control over personal information. As one participant asked, “Are the banks selling my information?” Trust is another major issue, with older adults fearing that AI systems may make errors or exhibit bias, particularly in sensitive domains such as healthcare and finance: “AI may give biased answers depending on who pays or who designs the system. AI recommendations could be unfair or influenced by location or tracking info. There is kind of dictatorship about how AI systems’ owners make decisions internally. Uncertainty around few people who are developing these systems.” Many also express anxieties about the loss of human contact, preferring that AI support complements rather than replaces face-to-face services: “Young people are losing the ability to socialise – are people going to depend on robots for companionship?” Complexity and usability remain significant barriers, as interfaces are often not designed with older users in mind, making technology difficult to navigate.
Despite these concerns, older adults perceive significant benefits and opportunities in AI. Technologies that support independence, such as smart reminders, home safety systems, and fall alerts, can help people live at home longer: “Continue to support activity of daily life, assist me in tasks (e.g. make my bed), stay in own home. Independence for as long as possible.” AI also offers potential for improved healthcare, including faster diagnosis, continuous monitoring through wearables, and access to remote consultations: “Personalised health apps to assist with day-to-day medical needs. Technologies that take into account the individual as a whole.” Tools that facilitate social connection, such as translation aids, voice assistants, and companion robots, help reduce isolation. Additionally, AI can simplify everyday tasks, from managing shopping lists to booking travel, often with simple voice commands or automated assistance. These opportunities suggest that when designed with older adults’ needs in mind, AI can enhance quality of life and foster autonomy.
Older adults are generally optimistic about AI when it is trustworthy, transparent, and easy to use. Many hope that AI can support independent living and improve daily routines without replacing the human interactions that are vital for well-being. There is strong interest in technologies that enhance engagement, learning and social connection, allowing older adults to remain active participants in family, community, and society. Participants also emphasized the value of education and skill-building: “More workshops, more training, having the discussion is very reassuring. Tutorship to address the gap in skills development,” and “There should be more courses delivered in an outreach way – available and accessible to all in the community (e.g. library).”
The paper has presented our co-creation methodology for engaging older adult stakeholders as part of the Age-Friendly AI project as well as initial findings from surveys and facilitated discussions. Responses from pre- and post-event surveys revealed significant improvements in participants’ reported perception of their understanding of AI and confidence in sharing their views about AI.
Facilitated discussions highlighted both challenges and opportunities in the adoption of AI by older adults. Concerns centered on privacy, security, trust, usability and the potential loss of human contact, yet participants also recognized clear benefits for independence, healthcare and social connection. Overall, older adults expressed cautious optimism, emphasizing the need for trustworthy, transparent and easy-to-use systems, alongside accessible education and training to build confidence and digital skills.
In the next phase of the project, these insights will form the basis of a series of multi-stakeholder workshops that bring together older adults, community organizations and policymakers. These sessions will create a structured space for dialogue, enabling stakeholders to critically reflect on the findings to date and to identify shared priorities and to build consensus on the design of an accessible AI literacy programme tailored for older adults. From the perspective of older adults, the programme will aim to equip them with the knowledge and skills needed to understand AI, make informed decisions, and safely engage with AI-enabled technologies. For community organizations, the initiative aims to contribute to capacity-building by providing transferable training materials that can be embedded into local community settings and existing digital inclusion activities. For policymakers, the project can yield insights into the enablers and barriers to AI literacy, offering a foundation for policy frameworks that support equitable age-friendly access.
The preferences shared to date by older adults are for a programme that adopts a balanced and nuanced approach, reflecting both the opportunities and challenges of AI. While older adults recognize the potential benefits of AI for daily life and broader societal applications such as healthcare, they also express concerns about trust and privacy. The programme will therefore aim to provide an honest and realistic view of AI, highlighting practical benefits while addressing potential risks and limitations. By doing so, participants can develop informed perspectives, make safer choices and engage confidently with AI technologies in ways that align with their values and priorities. Participants also indicated preferences for hands-on learning, outreach-based courses and skill-building workshops. The programme will incorporate these approaches, offering practical exercises, real-world examples and interactive discussions that allow participants to explore AI through different lenses, ethical, social and technical, such as through case studies and real-world scenarios, aiming to build both competence and confidence.
The programme will be delivered via a train-the-trainer model, leveraging trusted local networks such as community groups and the national library network. This approach will help to scale the initiative while also allowing participants to engage with it in familiar and supportive environments, ensuring accessibility, promoting trust and encouraging sustained participation. Ultimately, this approach seeks to enhance AI literacy, promote autonomy and empower older adults to navigate the evolving AI landscape with confidence and agency.
The research was conducted with financial support of the Research Ireland Discover programme under Grant No. 24/DP/1319 and the ADAPT Research Ireland Centre for AI-Driven Digital Content Technology under Grant No. 13/RC/2106_P2.
During the preparation of this work, the author(s) used ChatGPT in order to: grammar and spelling check, paraphrase and reword. After using this tool, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the publication’s content. [12] J. Jung, J. You, D. Kim, Effective but sustainable? A case of a digital literacy program for older adults, Educ. Inf. Technol. 30 (2025) 13309–13330. doi:10.1007/s10639-025-13364-4. [13] L.M.S. Miller, R.A. Callegari, T. Abah, H. Fann, Digital literacy training for low-income older adults through undergraduate community-engaged learning: single-group pretest-posttest study, JMIR Aging 7 (2024) e51675. doi:10.2196/51675. [14] M. Schirmer, K. Dalko, D. Stoevesandt, D. Paulicke, P. Jahn, Educational concepts of digital competence development for older adults—a scoping review, Int. J. Environ. Res. Public Health 20 (13) (2023) 6269. doi:10.3390/ijerph20136269. [15] B. Fischer, B. Östlund, N.K. Dalmer, A. Rosales, A. Peine, E. Loos, L. Neven, B. Marshall, Co-design as learning: the differences of learning when involving older people in digitalization in four countries, Societies 11 (2) (2021) 66. doi:10.3390/soc11020066. [16] A. Arola, M. Sandlund, M.E. Domellöf, et al., In their own words: older persons’ experiences of participating in co-creation, Res. Involv. Engagem. 11 (2025) 56. doi:10.1186/s40900-025-00725-z. [17] Research Ireland ADAPT Centre for AI-Driven Digital Content Technology.
https://www.adaptcentre.ie [18] L. Grehan, E. Clarke, C. Greene, 2021. #Discuss AI, ADAPT, the Research Ireland Centre for AI
Driven Digital Content Technology. https://www.adaptcentre.ie/discussai/ [19] Directorate-General for Research and Innovation (European Commission). 2025. Success stories from the research management community: a catalog of best practices and achievements. Technical Report. Publications Office of the European Union. https://data.europa.eu/doi/10.2777/7482920 [20] E. Clarke, C. Greene, L. Grehan, F. Murphy, Art or AI? A Museum of Very Modern Art, Proc.
The Public Communication of Science and Technology Conference (PCST 2025) (2025) p. 494.
Aberdeen, Scotland. https://www.abdn.ac.uk/events/conferences/pcst-2025/abstract-booklet/ [21] P. Kelly, E. Clarke, M. Parmar, D. Powell, M. Lynch, S. Ormonde, K. Colton, F. O’Meara, D.
O’Sullivan, D. Berry, Design and Development of an Interactive AI Toolkit for Engaging Older Adults in AI Discussions, Proc. 18th Int. Conf. PErvasive Technologies Related to Assistive Environments (PETRA ’25) (2025) pp. 285–291. doi:10.1145/3733155.3733217. [22] K. Ibrahim, S. Khodursky, T. Yasseri, Gender Imbalance and Spatiotemporal Patterns of Contributions to Citizen Science Projects: The Case of Zooniverse, Front. Phys. 9 (2021) Article 650720. doi:10.3389/fphy.2021.650720. [23] C. Jameson, V. Wylde, et al., Inclusive Approaches to Involvement of Community Groups in Health Research: The Co-produced CHICO Guidance, Res. Involv. Engagem. 9 (2023) Article 92. doi:10.1186/s40900-023-00492-9