IPILM is a learning environment that promotes collaborative knowledge construction among students from diverse cultural backgrounds. Educators and learners from various countries take part in an intercultural learning endeavor.
Embedded Youtube-Screencast: “IPILM 2025: AI and value propositions for stakeholder groups”.
Don’t have time to watch? No problem!
Session report and key aspects:
The presentation started with a brief introduction to AI and generative AI, followed by the section on roles of AI. There, three potential areas were introduced: customer experience, talent management and productivity, as well as risk management and governance.
AI-generated with ChatGPT
Particular attention was given to introducing the stakeholder groups: individual, organizational, and national and international stakeholders. The interest and power structure of each group was introduced, and showed their connectivity and interdependence.
AI-generated with ChatGPT
The perceived value of AI for individual stakeholders depends on life stage and occupational status, aligning with the distinct priorities of each group (UN Human Development Report 2025).
Used in the right way, AI may offer an opportunity to expand human capabilities. Institutional and social choices can enable AI to expand people’s capabilities and agency, as illustrated through AI’s applications for people with disabilities, care systems and gender equality, as well as in conceptualizing and mitigating AI bias. The following section on value propositions presented potential positive values.
AI-generated with ChatGPT
A part of the presentation also addressed several risks connected to AI. With a critical eye, we highlighted the risk of algorithmic bias, AI’s environmental impact, data privacy breaches, lack of transparency and cognitive debt.
AI-generated with ChatGPT
Discussion
This section addresses the key questions raised during the conference and summarizes the main points of discussion among participants.
The group was asked about their individual views on the personal value of AI tools. A variety of answers was given: while there were participants who did not view AI as particularly valuable, other participants saw great personal value in using the tools. Depending on the use cases and tasks, the majority saw a positive value.
Referencing the keynote speech on mindfulness, an open-ended question wanted the participants to share if they thought of AI as an inevitable feature and if humans would lose innate skills through the use of AI. This question sparked a discussion with different directions and focal points. The art of photography was named as an example for all three issues and sparked a dynamic discussion.
Another question wanted the participants to share their views on whether AI literacy should be introduced at an earlier age. The participants discussed that other types of literacy skills were introduced during childhood to prepare children at an early age and continually increase their skill set. The group members came to the conclusion that it should be similar with AI literacy education.
Conference Session Report: 7th IPILM-Conference on 11.12.2025
Original illustration, AI-generated with ChatGPT (OpenAI).
What happens when artificial intelligence becomes part of mental health care, and how should we deal with its risks and responsibilities?
Key Focus of the Session
AI as a support tool in mental health
Benefits: Early detection, accessibility, continuous support
Risks: Data protection, bias, transparency
Cultural and social contexts shaping perceptions, use, and risks of AI
Importance of information literacy and meta literacy
Building on these focal points, the Session examined the potential and limitations of artificial intelligence in the field of mental health from an information literacy and metaliteracy perspective. Drawing on a systematic literature review and concept mapping, it showed that AI-based applications can offer advantages, particularly with regard to early detection, continuous support, and low-threshold accessibility. These findings were largely consistent across the reviewed literature and were primarily informed by two key studies that shaped the session. Scientific evidence on the effectiveness and acceptance of AI-based mental health applications was mainly drawn from Dehbozorgi et al. (2025 – Read More). In contrast, ethical, cultural, and epistemic risks, such as data protection concerns, algorithmic bias, and limited transparency, were largely informed by the ethical review of Saeidnia et al. (2024 – Read more). The international survey largely reflected and reinforced the risks discussed in these studies, while also illustrating how these issues are perceived in practice. Overall, the findings emphasized that AI in mental health contexts should primarily be understood as a complementary tool to human expertise and that well-developed information literacy and metaliteracy are essential for responsible use.
❗️Below are a few selected examples of mental health services that incorporate AI-based support tools.
The examples illustrate current applications of AI in mental health and are not intended as recommendations.
Cultural and Ethical Considerations
Mental health is deeply shaped by cultural norms, social stigma, and structural inequalities, an aspect that was central to the intercultural perspective of the session and the conference as a whole. These factors also influence how AI-based systems are developed and used. AI applications risk reinforcing existing disparities through biased data, data poverty, and predominantly Western-centered models of mental health. Ethical challenges such as privacy, autonomy, and emotional adequacy are therefore particularly intensified for vulnerable and marginalized groups, highlighting the need for culturally sensitive and ethically grounded AI design.
Discussion
The discussion focused in particular on questions of responsibility. A majority of participants attributed responsibility for potentially harmful or misleading AI-based advice primarily to the providing companies, indicating a strong demand for institutional safeguards while simultaneously raising questions about the role of user responsibility. From an information literacy and metaliteracy perspective, this highlights the importance of enabling users to critically assess AI-based systems, understand their limitations, and recognize potential risks. At the same time, individual awareness alone cannot replace structural responsibility, especially in light of asymmetrical power and knowledge relations between providers and users, as well as the vulnerability of mental health contexts.
Another key point concerned the ambivalent level of trust in AI within mental health applications. Although many participants expressed general openness toward the use of AI, trust remained limited due to concerns about data protection, reliability, and the quality of AI-generated advice. Increasing trust was found to depend on transparent system design, strong data protection measures, explainable decision-making processes, and the clear integration of AI into human-supported care structures. Overall, the discussion suggests that trust in AI is shaped less by technological performance alone than by ethical design, cultural sensitivity, and informed and reflective practices of use.
Key Takeaways
AI can meaningfully support mental health care, but its value depends on ethical design, cultural sensitivity, and human oversight.
Users tend to view AI as a supportive tool rather than a replacement for professional care, while concerns about privacy and trust remain strong.
Cultural context plays a significant role in shaping how AI-based mental health services are perceived and used.
Strong information literacy and metaliteracy are essential for enabling critical, informed, and responsible engagement with AI in mental health contexts.
The following publications provide further insights into the scientific, ethical, and informational dimensions of AI in mental health contexts.
Dehbozorgi, R., Zangeneh, S., Khooshab, E. et al. The application of artificial intelligence in the field of mental health: a systematic review. BMC Psychiatry25, 132 (2025). https://doi.org/10.1186/s12888-025-06483-2
Li, H., Zhang, R., Lee, Y. C., Kraut, R. E., & Mohr, D. C. (2023). Systematic review and meta-analysis of AI-based conversational agents for promoting mental health and well-being. NPJ digital medicine, 6(1), 236. https://doi.org/10.1038/s41746-023-00979-5
Pellert, M., Lechner, C. M., Wagner, C., Rammstedt, B., & Strohmaier, M. (2024). AI Psychometrics: Assessing the Psychological Profiles of Large Language Models Through Psychometric Inventories. Perspectives on psychological science : a journal of the Association for Psychological Science, 19(5), 808–826. https://doi.org/10.1177/17456916231214460
Saeidnia, H. R., Hashemi Fotami, S. G., Lund, B., & Ghiasi, N. (2024). Ethical Considerations in Artificial Intelligence Interventions for Mental Health and Well-Being: Ensuring Responsible Implementation and Impact. Social Sciences, 13(7), 381. https://doi.org/10.3390/socsci13070381
AI can provide emotional comfort and a sense of social presence.
Users often perceive AI as non-judgmental and constantly available.
Emotional attachment to AI is possible, but true reciprocity is missing.
AI tends to replace functional roles rather than deep emotional relationships.
AI complements human relationships but does not replace them.
AI-generated image created using ChatGPT (DALL·E), 2026.
Introduction
This session aimed to provide an initial overview of the question of whether artificial intelligence can function as a substitute for human relationships. With AI tools increasingly used not only for information retrieval but also for emotional and social interaction, this topic has gained growing relevance in both academic research and everyday life.
The presentation combined a theoretical perspective, based on key findings from current scientific literature, with a practical approach. While the theoretical part outlined how human–AI relationships are conceptualized and evaluated in research, the practical component presented an exploratory survey to illustrate a possible research approach and to identify early tendencies in user experiences.
Summary
The presentation focused on two key academic studies and an exploratory survey to examine how AI may take on social and emotional roles traditionally associated with human relationships. The presented studies are: Brandtzaeg et al. (2022) showed that users candevelop friend-like attachments to social chatbots, perceiving them as safe and non-judgmental, while emphasizing the lack of reciprocity and emotional depth. Smith et al. (2025) further highlighted that although generative AI can convincingly simulate emotional responsiveness, it lacks key psychological components of genuine human connection, such as mutuality, shared experience, and emotional depth, which limits its ability to fully replicate human relationships. In addition, an exploratory online survey was conducted to demonstrate a possible research approach and to identify initial tendencies, such as emotional comfort, functional role substitution, and perceived non-judgment. Further details on the survey design, sample characteristics, and key findings are presented in the screencast linked below.
Discussion: Questions, Answers, and Reflections
During the discussion, a few questions focused on the methodology of the survey and the validity of its results. Participants critically addressed the small and non-representative sample. In response, it was emphasized that the survey was intended as an exploratory approach rather than a source of generalizable conclusions. Its purpose was to illustrate how human–AI relationships can be empirically examined and to reveal early tendencies that may guide future research. These included the frequent use of AI for emotional comfort, the perception of AI as less judgmental than humans, and the limited replacement of human roles.
Another discussion point concerned whether and how emotionally responsive AI systems should be regulated. It was debated whether emotional support provided by AI should be restricted and, if so, how “too emotional” AI could be defined. While arguments for regulation often focus on preventing emotional dependence, potential benefits are also emphasized, particularly AI’s role as a low-threshold form of support for individuals experiencing loneliness or social anxiety.
Finally, the discussion addressed broader opportunities and risks. Opportunities included availability, emotional relief, and reduced social pressure, whereas risks centered on privacy concerns, emotional dependence, and the potential weakening of real-life social relationships. Overall, the discussion underscored the need for continued critical reflection and interdisciplinary research on emotional AI.
Screencast
Below you can find the screencast of our presentation “AI as a Substitute for Human Relationships”, which summarizes the theoretical background and the practical insights discussed during the session.
Brandtzaeg, P. B., Skjuve, M., & Følstad, A. (2022). My AI friend: How users of a social chatbot understand their human–AI friendship. Human Communication Research, 48(3), 404–429. https://doi.org/10.1093/hcr/hqac008
Smith, M. G., Bradbury, T. N., & Karney, B. R. (2025). Can generative AI chatbots emulate human connection? A relationship science perspective. Perspectives on Psychological Science, 20(6), 1081–1099. https://doi.org/10.1177/17456916251351306
Hohenstein, J., Kizilcec, R. F., DiFranzo, D., Aghajari, Z., Mieczkowski, H., Levy, K., & Jung, M. F. (2023). Artificial intelligence in communication impacts language and social relationships. Scientific Reports, 13, 5487. https://doi.org/10.1038/s41598-023-32354-5
Malfacini, K. (2025). The impacts of companion AI on human relationships: Risks, benefits, and design considerations. AI & Society. https://doi.org/10.1007/s00146-025-02318-6
Zimmerman, A., Janhonen, J., & Beer, E. (2024). Human/AI relationships: Challenges, downsides, and impacts on human/human relationships. AI and Ethics, 4, 1555–1567. https://doi.org/10.1007/s43681-023-00348-8
Artificial intelligence is increasingly shaping political communication and democratic processes. Generative AI systems such as deepfakes, automated text generation, and AI-supported political campaigning are transforming how political information is produced, disseminated, and perceived. The screencast “AI and Political Dis- and Misinformation” examined these developments from a comparative perspective, combining current research, international case studies, and exploratory empirical observations.
This landing page summarizes the key arguments and insights presented in the screencast and discussed during the session.
Screencast
The following screencast provides an overview of current research on AI-driven political misinformation, comparative case studies from different national contexts, and exploratory empirical findings discussed during the conference.
Drawing on recent research in political science and media studies, the screencast situated AI-driven misinformation within broader sociopolitical debates about power, manipulation, and democratic stability. Scholars emphasize that AI does not merely accelerate existing forms of disinformation, but qualitatively transforms political propaganda by increasing its scale, personalization, and plausibility (Gaborit, 2024; Romanishyn et al., 2025).
Particular concern has been raised about the capacity of generative AI to undermine trust in political information ecosystems, especially during election periods. Public anxieties around AI and misinformation are often shaped not only by concrete incidents, but also by media narratives and broader fears about democratic erosion (Yan et al., 2025).
Comparative Case Studies
To ground these debates empirically, the screencast presented three case studies focusing on recent electoral contexts in the United States, India, and Germany.
United States
The U.S. case focused on the 2024 presidential election and the origins of public concern about AI “supercharging” political misinformation. Research suggests that fears of AI-driven manipulation were strongly influenced by public discourse and media coverage, even in cases where documented AI misuse remained limited (Yan et al., 2025). This highlights the importance of perception, trust, and anticipatory regulation in democratic contexts.
India
The Indian case demonstrated more direct forms of AI misuse in political communication. Generative AI tools, including deepfakes and manipulated audiovisual content, were actively deployed during the 2024 elections, contributing to misinformation, voter confusion, and political polarization. These developments illustrate both the technological possibilities and democratic risks of AI in highly mediated political environments (Dhanuraj et al., 2024).
Germany
The German case examined AI-based voter information tools developed ahead of the 2025 federal elections. Although intended to provide neutral political guidance, these systems sometimes produced biased or misleading outputs. This case serves as a cautionary example of how “neutrally” informative AI tools can unintentionally become sources of political misinformation, raising questions about transparency, accountability, and design assumptions (Dormuth et al., 2025).
Together, the case studies show that AI-driven political misinformation manifests differently across national contexts, but consistently challenges democratic trust and decision-making.
Exploratory Survey Insights
In addition to the case studies, the screencast presented findings from an exploratory online survey conducted in November 2025 with 108 participants from India, the United States, and Germany. The survey examined awareness of AI-generated political content, experiences with political misinformation, trust in political information on social media, and attitudes toward regulation and responsibility.
Across all three countries, respondents reported frequent exposure to political misinformation and relatively low confidence in their ability to identify AI-generated fake content. Trust in political information shared on social media platforms was generally low. These findings are tentative and illustrative, and primarily serve to contextualize the case studies rather than to provide representative conclusions.
Discussion and Ethical Implications
The discussion following the screencast addressed key normative and practical tensions. A central debate concerned prevention versus detection: whether democratic responses should focus on restricting and labeling AI-generated political content or prioritize detection mechanisms and citizen awareness.
Closely related was the question of regulation and education. Participants emphasized that regulation and AI literacy should not be understood as mutually exclusive, but rather as complementary strategies. Given the increasing sophistication of generative AI systems, even digitally literate users remain vulnerable, underscoring the need for combined policy and educational approaches.
Further discussions addressed platform and developer responsibility, the risk of bias in AI systems, and the limits of existing legal frameworks in addressing AI-driven electoral manipulation.
Conclusion
The session and screencast demonstrated that AI-driven political misinformation represents a serious and evolving challenge for democratic societies. While national contexts differ, the underlying issues of trust, transparency, and accountability are shared across political systems.
Addressing these challenges requires interdisciplinary responses that combine regulatory frameworks, responsible AI design, platform accountability, and public AI literacy. As generative AI continues to develop, proactive and ethically informed strategies will be essential to safeguard democratic communication.
References
Dhanuraj, D., Harilal, S., & Solomon, N. (2024). Generative AI and its influence on India’s 2024 elections: Prospects and challenges in the democratic process. Friedrich Naumann Foundation for Freedom.
Dormuth, I., Franke, S., Hafer, M., Katzke, T., Marx, A., Müller, E., & Rutinowski, J. (2025). A cautionary tale about “neutrally” informative AI tools ahead of the 2025 federal elections in Germany. In Proceedings of the World Conference on Explainable Artificial Intelligence (pp. 64–85). Springer Nature Switzerland.
Gaborit, P. (2024). A sociopolitical approach to disinformation and AI: Concerns, responses and challenges. Journal of Political Science and International Relations, 7(4), 75–88.
Romanishyn, A., Malytska, O., & Goncharuk, V. (2025). AI-driven misinformation: Policy recommendations for democratic resilience. Frontiers in Artificial Intelligence, 8, 1569115.
Yan, H. Y., Morrow, G., Yang, K. C., & Wihbey, J. (2025). The origin of public concerns over AI supercharging misinformation in the 2024 US presidential election. Harvard Kennedy School Misinformation Review.
This session was part of the course Intercultural Perspectives on Information Literacy and Metaliteracy (IPILM, Winter Semester 2025/26) and was presented at the IPILM Conference on December 19, 2025. It examined how artificial intelligence is transforming the production of digital media and which ethical challenges arise from this development. The focus was on authorship, responsibility, bias, trust and the future role of human creativity.
🎥 Screencast
The contribution was presented as a screencast designed as an Open Educational Resource (OER). The screencast combines scientific research with practical insights from artists and case studies.
The discussion focused on the future of art in the age of AI:
Will AI replace human-made art or function as an additive tool, similar to photography?
Who bears responsibility for ethical AI use: artists, developers, platforms, or regulators?
How can audiences distinguish AI-generated from human-created content?
Emerging technical solutions such as C2PA (Coalition for Content Provenance and Authenticity) standards were referenced as potential mechanisms to support transparency and verifiable content provenance.
📄 Session Report
A detailed written session report is available here:
Landingpage for conference session on the IPILM blog: 7th IPILM-Conference
Published by Fabienne Katharina Müller
AI-generated with chatgpt
Quick facts and insights
Role of AI
AI broadens access via adaptive tutoring, predictive analytics, and multilingual support (Yeo & Lansford, 2025).
ITS (Intelligent Tutoring Systems) demonstrate measurable improvements in learning outcomes in diverse contexts (Holmes et al., 2019).
UNESCO (2024) stresses AI’s potential in low‑resource environments when supported by policy.
What are important questions?
What is AI’s role in improving access to education for diverse learners?
How does AI help in personalised learning, and why is it important for inclusive education?
In what ways is AI transforming the job market and creating new opportunities?
What skills do learners need to stay relevant in an AI-driven job environment?
How can AI support equal access to job information and career guidance?
Summary of the topic
This blog examined the role of artificial intelligence in improving access to education and its broader implications for the job market. A key focus was on how AI can support inclusive education through personalized learning, intelligent tutoring systems, adaptive feedback, and multilingual support, thereby addressing diverse learning needs. At the same time, the presentation critically discussed structural and ethical challenges, including algorithmic bias, data protection and privacy risks, limited transparency of AI systems, and generally low levels of AI literacy among users. In addition, the presentation highlighted global inequalities in access to AI, emphasizing that countries with stronger digital infrastructure and higher AI preparedness benefit more from AI adoption, while others risk being left behind. Two empirical case studies were used to support these points: one analyzing teachers’ trust in AI in education across different countries, and another examining the impact of generative AI on employment, skill requirements, and labor market inequalities. Overall, this emphasized that AI offers significant opportunities, but only if implemented responsibly, ethically, and with equal access in mind.
AI-generated with chatgpt
🟢Advantages
of AI in Job Market (Cazzaniga et al. 2024), (UNESCO 2024),(Holmes et al. 2019)
AI enhances task efficiency for high-skilled workers and increases productivity, allowing them to focus on strategic and creative tasks.
AI creates hybrid roles that integrate human judgment with AI-assisted decision-making, expanding job categories in data analytics, automation management, and AI oversight.
AI-powered job platforms improve matching accuracy by analysing skills, experience, and job requirements, supporting equal access to opportunities.
Countries withhigh AI preparedness gain new employment pathways through digital infrastructure, training, and innovation ecosystems.
AI adoption pushes workers to acquire modern digital skills, improving long-term employability and adaptability.
🔴Disdvantages
of AI in Job Market (Cazzaniga et al. 2024)
AI-driven automation replaces repetitive, low-complexity tasks, increasing unemployment risk for low-skilled and ageing workers.
Differences in AI readiness create unequal job opportunities globally, widening economic and employment gaps.
AI may disproportionatelyaffect women and certain professional groups, increasing vulnerability to job displacement.
AI leads to wage inequality, where high-skilled workers gain from productivity increases while low-skilled workers face stagnant or declining wages.
Workers displaced by AI find it harder to transition into new roles without advanced digital skills, increasing long-term unemployment risks.
The video shows why job losses are occurring in some areas, which professions are particularly affected, and which skills will be crucial in the future. It also discusses how to strategically develop in your current job, the continuing role of education, and why building a strong personal positioning is becoming increasingly important in the age of AI. Finally, a clearly structured three-step approach is presented for remaining professionally relevant in the long term. https://www.youtube.com/watch?v=l6d_0PB0Pbg
🟢Advantages of AI in Education
(UNESCO, 2024), (Yeo & Lansford, 2025), (Holmes et al., 2019)
Healthcare
AI simulations allow students to practise surgeries and diagnoses safely.
Predictive models help students understand real-world medical decision-making.
Multilingual virtual assistants support global medical learners.
Finance
AI financial modelling tools prepare students for real-market scenarios.
Al Khan, founder of Khan Academy, is convinced that artificial intelligence can greatly improve the education system. He shows how AI can support students through personalized learning assistance and teachers through digital assistance systems, and introduces new features of the educational chatbot Khanmigo. https://www.youtube.com/watch?v=hJP5GqnTrNo
In the healthcare sector, bias, data protection issues, and a lack of transparency can lead to incorrect or unfair AI decisions, while a lack of human contact and low AI literacy further complicate care.
Finance
In the financial sector, bias, data protection risks, and non-transparent AI models have a significant impact on fairness and trust, especially when professionals lack AI expertise.
Education
In education, the disadvantages of AI mainly concern academic integrity, data protection, algorithmic fairness, lack of human support, and generally low AI literacy.
Key findings from two relevant case studies
Case Study 1 – Job Market “Gen-AI: Artificial Intelligence and the Future of Work.” (Cazzaniga et al. 2024)
Highly skilled jobs are most affected by AI, but also benefit the most (increased productivity, better wages).
Low-skilled and older workers are at greatest risk of being disadvantaged by AI.
Women and knowledge workers are particularly exposed to AI.
AI can exacerbate inequalities, especially in countries with poor digital preparedness.
The US/UK are well prepared, emerging markets less so, resulting in large global differences.
Case Study 2 – Education “What Explains Teachers’ Trust in AI in Education Across Six Countries?“(Viberg et al., 2025)
Perceived benefits ↑ → Trust ↑; Concerns ↑ → Trust ↓. These two were the strongest predictors of trust.
AI self-efficacy & AI understanding strongly increased perceived benefits and reduced concerns — indirectly boosting trust.
Demographics (age, gender, education) did not significantly influence trust.
Cultural values mattered: High uncertainty avoidance, collectivism, and masculinity were associated with differences in trust and concerns.
Cross-country variation: Brazil, Israel, and Japan showed higher trust; Norway, Sweden, and USA showed lower trust after adjustments.
🗣️Discussion
Which aspects stood out particularly in relation to this topic?
Al-Zahrani, A.M., Alasmari, T.M. (2024): Exploring the impact of artificial intelligence on higher education: The dynamics of ethical, social, and educational implications. Humanit Soc Sci Commun 11, 912 https://doi.org/10.1057/s41599-024-03432-4.
Cazzaniga et al. (2024): “Gen-AI: Artificial Intelligence and the Future of Work.” IMF Staff Discussion Note SDN2024/001, International Monetary Fund, Washington, DC. https://doi.org/10.5089/9798400262548.006 .
Marín, Y. R., Caro, O. C., Rituay, A. M. C., Llanos, K. A. G., Perez, D. T., Bardales, E. S., Tuesta, J. N. A. & Santos, R. C. (2025): Ethical Challenges Associated with the Use of Artificial Intelligence in University Education. Journal Of Academic Ethics, 23(4), 2443–2467. https://doi.org/10.1007/s10805-025-09660-w .
Sahar, R. & Munawaroh, M. (2025): Artificial intelligence in higher education with bibliometric and content analysis for future research agenda. Discover Sustainability, 6(1). https://doi.org/10.1007/s43621-025-01086-z .
UNESCO. (2024): AI and inclusive education: Policy guidance for promoting equity. United Nations Educational, Scientific and Cultural Organization. https://doi.org/10.54675/PCSP7350 .
Viberg, O., Cukurova, M., Feldman-Maggor, Y., Alexandron, G., Shirai, S., Kanemune, S., Wasson, B., Tømte, C., Spikol, D., Milrad, M., Coelho, R. & Kizilcec, R. F. (2025): What Explains Teachers’ Trust in AI in Education Across Six Countries? International Journal of Artificial Intelligence in Education, 35, 1288–1316. https://doi.org/10.1007/s40593-024-00433-x.
Yeo, G. & Lansford, J. E. (2025): Effects of Artificial Intelligence on Educational Functioning: A Review and Meta-Analysis. Educational Psychology Review, 37(4). https://doi.org/10.1007/s10648-025-10085-5 .
