Foresight Methodologies in Responsible GenAI Education: Insights from the Intermedia-Lab at Complutense University Madrid

Abstract

This study, conducted at Complutense Intermedia-Lab, employs a dual approach to explore university students’ use of Generative AI (GenAI), combining a survey with foresight methodologies (Sci-fi prototyping). The initial survey gathers baseline data on students’ experiences, attitudes, and concerns regarding GenAI, providing a comprehensive understanding of current practices among university students in Spain. This empirical foundation informs subsequent Sci-fi prototyping sessions, where students creatively envision future scenarios, fostering futurist thinking and deeper engagement. By integrating principles of Responsible Research and Innovation (RRI), this approach facilitates a nuanced exploration of GenAI’s potential impacts on education. The incorporation of both quantitative data collection and qualitative foresight methods in this study serves to navigate challenges and level opportunities of promoting the ethical and inclusive incorporation of GenAI in Higher Education, ensuring that future innovations align with societal values and needs.

1. Introduction

Effective foresight has always been crucial in human life, a vital skill for survival. It implies a very human trait, the ability to anticipate and plan for the future. Today, technological innovation is happening at a pace that can be disorienting. In such a dynamic environment, characterized as VUCA (acronym for Volatility, Uncertainty, Complexity, and Ambiguity), the ability to anticipate and prepare for future developments becomes even more important. In this scenario, it is easy to succumb to a fatalistic mind-set, but fostering foresight can help proactive thinking and resilience. By thinking ahead, individuals can make strategic plans and informed decisions, reducing the likelihood of being caught off guard by unexpected rapid changes.

The implementation of digital technologies in the 1990s brought significant shifts in innovation processes within institutions, the job market, society in general, and education environments in particular. More recently, the onset of AI is bringing about even more radical transformations. Generative Pre-trained Transformers like ChatGPT, Dall-E, Midjourney, etc., are causing far-reaching transformations in all aspects of our daily lives, including education. Technology is more than ever a strategic resource, and it is important to anticipate and prepare for its future impact and opportunities.

In the initial stage of implementing new technologies, there are both fears and expectations, and there can be a mismatch between the speed of their introduction and their social relevance. It cannot be expected that innovations requiring transformative infrastructural changes will be rapidly implemented. The consensus of key stakeholders, who play a significant role in the governance of innovations as well as their planning, allocation, and promotion, is fundamental. This planning also requires the inclusion of aspects of responsible fair use.

Responsible Research and Innovation (RRI) was first introduced as a concept in the European Union’s (EU) research and innovation framework in the early 2010s. It gained prominence with the launch of the Horizon 2020 program (2014–2020), with RRI principles integrated into various funding schemes and projects, marking a significant step towards embedding responsible practices, including aspects of ethics, gender equality, public engagement, and open access. Since then, RRI has continued to evolve and expand its influence in EU research and innovation policy, evolving to include Artificial Intelligence (AI) by recognizing the need to address its ethical, social, and environmental implications. A Group of Experts was created and a report published, setting out key principles that should guide the development and use of AI technologies [1].

The European RRI perspective includes social responsibility as a quality dependent on the inclusion of a plurality of actors in any innovation dynamics. Thus, the teaching of RRI has been introduced in all the levels of education programs throughout the EU. In recent years, RRI has begun to include aspects directed to orienting students about the responsible use of GenAI that may have ontological and epistemological consequences for the future. RRI principles have been included in the survey conducted in the study presented in this article, aimed at identifying the challenges that the use of GenAI among students has introduced in Higher Education institutions in Spain. Advances in GenAI introduce new dimensions in understanding RRI and academic responsibility, moving beyond reliance solely on scientific evidence and encompassing aspects such as co-creation and accountability for results generated by means of GenAI.

Based on results from an initial survey, the research presented involves the use of foresight methodologies in the context of Philological Studies in general, and English Studies at Complutense University Madrid in particular, in order to assess the impact of GenAI in Higher Education in Spain. Although the sample is not as ambitious as we would have liked, it still offers some guiding background on the concerns surrounding the use of GenAI. In order to test and extend some of these results, foresight methodologies were introduced at a later stage in the project. These methodologies are also aligned with RRI, not only because they encourage academic diligence and responsibility. Foresight also involves the cultivation of a mind-set that values long-term planning and positive adaptability in a rapidly evolving world, teaching students to anticipate future trends and enabling them to make informed decisions about their education and career paths. Incorporating foresight training into educational curricula can empower students to navigate uncertainties, such as those unveiled in the survey, develop resilience, and seize emerging opportunities.

The initial impression is that AI learning systems can hold the promise of a more personalized instruction, adapted to individual student needs and learning styles providing customized feedback. These systems might also assist teachers in data analytics, improving tracking of student progress, tutoring and grading, helping with consistent assessments, and the optimization of resource allocation. Additionally, AI translation tools can facilitate the breakdown of language barriers in multicultural educative contexts. Overall, these tools contribute to more accessible and flexible education, including blended learning models and the accommodation of diverse learning preferences. These technologies might also provide more engaging and immersive learning experiences through interaction, gamification, and even virtual reality. Finally, they might enable opportunities for ongoing education and professional development, supporting lifelong learning initiatives as well as enhancing skills that are relevant for future employment in certain professional fields.

This apparently positive scenario can also bring significant challenges, among them, ethical concerns around issues such as GenAI being trained on biased data that could perpetuate and even amplify existing social inequalities, leading to unfair treatment of certain groups. There is also apprehension about the privacy and security of information, as well as fears that AI could lead to job displacement for some educational roles. Finally, over-reliance on GenAI in education can lead to the loss of originality in students’ ideas as well as a loss of personal writing style, to the point of compromising the authenticity of students’ performances and causing unintentional plagiarism arising from overuse. Balancing the benefits and challenges of GenAI is crucial in order to foster RRI innovation while safeguarding societal values and human well-being.

The article begins with a brief overview of foresight methodologies and their role in education. Two different foresight approaches are introduced in the context of the research carried at the Intermedia-Lab of the Department of English Studies, Faculty of Philology, Complutense University Madrid. The first method is quantitative, by means of a survey. Surveys are not typically classified as foresight methods per se, but they can play a crucial role within the broader context of foresight activities. The second activity employs Sci-fi prototyping to explore future avenues for the use of AI in education in order to pinpoint some of the challenges faced from a qualitative perspective. Using a survey to examine university students’ use of GenAI offers a foundational understanding of their experiences, attitudes, and concerns, providing critical baseline data and ensuring a comprehensive grasp of current practices and perceptions. Following the survey, Sci-fi prototyping allows students to creatively envision future scenarios, fostering innovative thinking and deeper engagement. The initial survey informs the Sci-fi prototyping, ensuring that the speculative narratives are grounded in real-world insights. This combination enables a nuanced exploration of GenAI’s potential impacts on education, blending empirical data with imaginative foresight to anticipate and shape future developments.

1.1. Foresight Methodologies

Foresight methodologies were first implemented in the mid-20th century, originating from strategic military planning during World War II and the Cold War. The RAND Corporation in the United States was a pioneer think tank in the late 1940s and 1950s, developing techniques such as scenario planning and the Delphi method to anticipate future technological and military developments. The concept expanded beyond military applications in the 1960s and 1970s, entering government policy, corporate strategy, and academia [2,3]. In the 1990s, foresight had become an established practice in many countries, with national and international foresight programs being launched [4,5,6]. The European Union, for example, began incorporating foresight into its research and innovation policies, leading to the development of various foresight initiatives to address societal challenges and inform strategic planning, including education [7] and RRI projects such as NewHoRRIzon, in which the author of the article took part [8].

The prevalence of a foresight method often depends on the context and the goals of the organization or project, which can include strategic planning, innovation, and creativity to prototype future scenarios, or the ability to gather expert opinions and identify emerging issues in policy-making [9]. The most widely used foresight methods include the following:

• Horizon scanning: This involves an ongoing effort to recognize significant changes occurring outside the purview of the organization or group engaging in the scanning process. Typically, it entails systematically surveying newspapers, magazines, websites, and other media to identify trends that are indicative of future developments [10].
• Historical analysis: Comparing current situations to past events helps inform decision-making processes based on historical insights [11].
• Trend analysis: This method involves identifying and analysing trends over time to forecast future developments, underlying causes, speed of progression, and potential impacts and implications. This method often relies on historical data and statistical techniques that help extend this trend line into the future based on recent rates of change [12].
• Scenario development: Scenarios are outlines that depict plausible future possibilities based on existing knowledge or assumptions. This method involves narrative and storytelling techniques [13].
• Delphi method: This involves a structured framework where a panel of experts provides insights and forecasts through methods such as face-to-face discussions, online or telephone interviews, or surveys and questionnaires circulated among participant groups [14].
• Models: These are representations of real-world phenomena that facilitate a deeper understanding of complex systems, for example, architectural models for future buildings, maps of geographic areas, etc. They can be simulated using computer models [15].
• Simulations or Gaming: These involve dynamic representations of real-world scenarios, often used to test alternative strategies and tactics [16].
• Brainstorming: This method involves generating new ideas through collaborative group discussions aimed at exploring creative solutions or opportunities by encouraging participants to build on each other’s ideas, fostering idea generation and co-creation.
• Science Fiction Prototyping: This involves exploring Science Fiction and Speculative Narratives to visualize future possibilities [17,18].

1.2. Foresight in Education

Foresight methodologies have been increasingly used in education to anticipate and shape future trends, challenges, and opportunities. For instance, they have been applied in the educational sector to anticipate future job market skills and knowledge requirements, leading to relevant and forward-looking curricula as well as helping institutions in creating educational strategies and policies that align with future societal needs [19]. The implementation of RRI in education can benefit from these methodologies, where interventions often include collective dynamics based on Design Thinking and Problem-Solving approaches [20]. Prospective methodologies have also contributed to identify emerging research areas and keep academic institutions at the forefront of innovation and societal challenges [21].

Some projects that have made use of foresight methodologies are the following:

• The Future Classroom Lab project, initiated by European Schoolnet. It employs foresight methodologies to envision future classroom environments and develop strategies for integrating technologies such as augmented reality, virtual reality, and collaborative platforms into teaching practices [22].
• The Horizon Report produced by the New Media Consortium (NMC) and the EDUCAUSE Learning Initiative also uses foresight methods to identify emerging technologies with potential impacts on education, including artificial intelligence, blockchain, and immersive learning environments [23].
• The Education 2030 project led by UNESCO explores scenarios for how education systems may evolve in response to technological advancements, demographic shifts, and global challenges [24].
• Foresight4EU, funded by the European Commission, applies foresight methodologies to explore the potential impact of emerging technologies on various sectors, including education. It engages stakeholders in scenario-building exercises to anticipate how technologies like artificial intelligence, robotics, and big data analytics may transform teaching and learning processes in the future [25].
• The Schooling of Tomorrow Initiative, led by the OECD, employs foresight to examine future trends in education and identify innovative practices for school transformation, including adaptive learning systems, personalized learning platforms, and learning analytics [26].
• AI4EU is a European project aiming to build a European AI ecosystem, including applications in education [27].
• The EDUCAUSE Learning Initiative (ELI) explores the intersection of AI and education through articles, webinars, and resources aimed at Higher Education institutions. It discusses the potential of AI technologies to enhance teaching, learning, and institutional effectiveness [28].
• The European AI Alliance engages stakeholders across various sectors, including education, to discuss AI-related opportunities and challenges. While not dedicated solely to education, it addresses AI applications and implications in education among other areas [29].

1.3. Science Fiction Prototyping

Science fiction prototyping was developed by Brian David Johnson at Intel [30]. The method involved creating short pieces of fiction, known as science fiction prototypes—stories, movies, or comics—that potentially envision technological innovations and their potential implications for future societies, including the social, ethical, and practical consequences of emerging technologies. Johnson’s method comprised five sequential stages, as seen in

Table 1. Brian David Johnson’s Steps in Science Fiction Prototyping.

Stage	Description	Purpose
Identify the Technology	Select the technology and underlying science to be explored. This could be an emerging technology or a recent scientific discovery	Establish the focal point of the science fiction prototype, serving as the foundation for the narrative.
Esplore Future Implications	Speculate on the future implications of the chosen technology for society. Consider how it might improve or worsen people’s lives, potential risks, and new challenges	Understand the broader impact of the technology on society and identify key issues to address the narrative.
Create the narrative	Develop a story that conveys the answers to the questions raised in the previous step. This involves thinking about the characters, their experiences, and the technology’s effects on their lives.	Use storytelling to bring the speculative future to life, making it more tangible and relatable for the audience.
Build the Prototype	Translate the narrative into a science fiction prototype, which could be a written story, a comic, a movie, or another form of media.	Create a concrete representation of the future scenario that can be shared and discussed with others.
Reflect and Learn	Use the story as a tool for learning and reflection. Consider what could have been done differently to prevent conflicts or negative outcomes depicted in the narrative.	Extract lessons and insights that can inform present-day decisions and strategies related to the technology.

below.

The Intermedia-Lab for English Studies at Complutense University Madrid has been using foresight methodologies in order to help them envision the widest possible range of career paths and entrepreneurship [31]. In the case of the research presented in this article, Sci-fi has been used to envision the future of GenAI in Higher Education and professional careers. Watching or reading Sci-fi works as part of the syllabus in narrative fiction helps increase emotional involvement with technological changes taking place, stimulating ideas about coming times. By analysing these narratives, participant students can generate alternative scenarios and explore the ethical, social, and philosophical implications of GenAI technologies, informing their prototyping efforts after collaborative Design Thinking and Problem-Solving activities oriented towards RRI [32].

In the last two decades, there have been numerous explorations of Sci-fi prototyping to forecasting technological and social change [33,34]. Some have focused on the role of context [35]. Others have explored scenario archetypes [36]. In 2020, Alessandro Fergnani and Zhaoli Song published the results of their examination of a sample of 140 Science Fiction films in order to identify common patterns or themes and group them into a distinct scenario of six predetermined future scenarios [37]. They also considered the dimensions of STEEP (Social, Technological, Economic, Environmental, and Political) to facilitate implications across dimensions. They came up with the following categories: a “Growth & Decay” scenario that reflects capitalist structures where corporations hold unchecked power, leading to technological advancement driven by profit. Common themes include social decay, disparities, environmental degradation, and moral decline. Films like “Metropolis”, “Blade Runner”, and “Avatar” illustrate these elements, highlighting societal issues like social disconnection and ethical dilemmas in AI, as seen in “Her” and “Ex Machina”. Catastrophic events often lead to societal instability, crime, and military control, as depicted in “Idiocracy”, “Dredd”, and “Children of Men”. In the “Threats & New Hopes” scenario, humanity faces imminent catastrophic threats, prompting global cooperation to prevent disaster, with themes of sacrifice and unity central to films like “2012”, “Pacific Rim”, “Transcendence”, and “Interstellar”. The “Wasteworlds” scenario depicts post-catastrophic events resulting in harsh living conditions and societal regression, focusing on survival, as seen in “Mad Max”, “Waterworld”, “The Postman”, and “WALL-E”. “The Powers that Be” scenario features totalitarian regimes emerging after catastrophic events, controlling technology and suppressing freedoms, with resistance as a key theme in films like “Aeon Flux”, “Hunger Games”, “Divergent”, and “The Giver”. In the “Inversion” scenario, humanity is dominated by superior entities, such as aliens, depicted in “A Quiet Place”, “After Earth”, “Planet of the Apes”, and “Alien: Covenant”. Finally, the “Ethical and Moral Implications” scenario explores autonomy, consciousness, and rights of sentient beings, with themes of AI rebellion and the nature of consciousness, as exemplified in “I, Robot”, “AI: Artificial Intelligence”, “Westworld”, “Blade Runner 2049”, “Ghost in the Shell”, and “Transcendence”.

Drawing inspiration from Fergnani and Song’s research, the project developed at Complutense Intermedia-Lab aimed at identifying similar trends focusing on the impact of GenAI in the future of education.

2. Materials and Methods

Students of English Studies and Philology typically have an interest in languages, literatures, and world cultures. Their academic journey encompasses a comprehensive study of linguistics—phonetics, syntax, semantics, and pragmatics—literary analysis, and the historical evolution of literary genres, periods, and movements, cultural studies, as well as critical theory. Students of Philology develop strong analytical and interpretative skills, enabling them to dissect complex texts and understand diverse literary traditions and cultural contexts. Additionally, they cultivate proficiency in written and oral communication, essential for articulating nuanced arguments and engaging in scholarly discourse. This multidisciplinary approach equips them with the ability to analyse language and literature from multiple perspectives, so that the avenues for their future employment are diverse and multifaceted. Many pursue careers in education, becoming teachers or lecturers in schools, colleges, and universities, dedicated to inspiring a love of reading and critical thinking in their students. Others find opportunities in the publishing industry, working as editors, proofreaders, or literary agents, where their skills in language and attention to detail are invaluable. They may also work in communications, public relations, translation, and interpretation, where their ability to articulate ideas clearly and persuasively is highly sought after.

A media lab is an interdisciplinary research and development facility focused on the intersection of technology, media, and creative practices. Typically associated with academic institutions or innovation centres, media labs bring together experts from interdisciplinary fields, including sciences, social sciences, humanities, and the arts, to collaborate on cutting-edge projects. These labs conduct research in areas such as digital media, AI, virtual reality, and human–computer interaction, developing new technologies and methodologies. They provide a space for creative experimentation, allowing for the prototyping and testing of innovative ideas. Media labs also serve as educational hubs, offering courses and hands-on training to students and professionals, and support entrepreneurial activities by assisting start-ups with resources and mentorship. Additionally, they engage with knowledge-transfer and Open Science activities to the general public through exhibitions and events, disseminating research findings and fostering discussions about emerging technologies. Overall, media labs advance media technology, foster education, and drive digital innovation. The Complutense Intermedia-Lab was put in motion as part of the knowledge-transfer initiatives of the research group Studies on Intermediality and Intercultural Mediation (SIIM), dependent on the Department of English Studies at the Faculty of Philology [38]. SIIM Intermedia-Lab activities have encompassed a broad range of initiatives designed to disseminate and extend university students’ knowledge, expertise, and skills to the real world, and fostering collaboration across various social sectors. They have included training programs, conferences, workshops and seminars, collaborative projects, publications, and reports.

2.1. Identifying Challenges

In the case of research presented in this study, we first used a survey to gather foundational data on university students’ experiences, attitudes, and concerns regarding the use of GenAI in their educational experiences. The survey provided critical baseline data. Subsequently, Sci-fi prototyping allowed students to imagine future scenarios.

The initial survey was created using Google forms in order to assess students’ level of awareness of the benefits as well as potential problems of using GenAI models in their classroom experiences [39]. Some of the questions were based on the UNESCO (2023) guide to the use of ChatGPT and AI in Higher Education, fostering RRI principles [40].

The questionnaire also incorporated a final list of key takeaways to bear in mind for a responsible use of these tools in education environments. The survey was open during the duration of the course in 2023–24 and collected 72 responses. Findings can be visualized in the following multi-section charts, including bar charts, highlighting the frequency of use (daily, weekly, occasionally, never) and the percentage impact for different academic tasks. Pie charts show the overall percentage of respondents who believe AI has no impact versus those who see specific benefits. The majority of students pursuing English Studies are female, so it is not surprising to see their rate of response in comparison to other gender groups. Students’ ages are diverse, though mostly in their early twenties (Figure 1).

The survey also aimed to gain insights into students’ future career aspirations to determine the optimal ways to integrate AI into their learning experience. The majority envisioned careers in private enterprises or national government-owned companies. Notably, 18% aspired to become university professors—a concerning figure given the limited opportunities and competitive nature of the academic sector in Spain (Figure 2).

When asked about their frequency of using GenAI systems (Figure 3), it was surprising to discover that most students reported either never using them or using them only occasionally, primarily for entertainment purposes. Despite the survey’s anonymity and confidentiality, these responses may indicate a level of distrust or hesitancy towards using these technologies in an academic context.

Regarding the functionalities of ChatGPT, for instance, 54% of students claimed not having any knowledge about its customisation (Figure 4).

While most students reported not using GenAI systems in their academic pursuits, a different picture emerged when asked about the impact of these systems on their classwork. Interestingly, responses indicated that GenAI could assist in understanding concepts, brainstorming, and idea generation. More than 30% of students believed that it enhanced their overall writing style (Figure 5). Based on the data presented, it is evident that although a majority of students reported not using GenAI systems in their academic environment, a significant proportion believed that these systems had a positive impact on their classwork. Specifically, over 30% of respondents felt that GenAI improved their overall writing style. This discrepancy between reported usage and perceived impact suggests that while students may not actively engage with these systems, they recognize their potential benefits in enhancing academic tasks such as understanding concepts, brainstorming, and generating ideas.

The chart reveals a cautious and selective adoption of GenAI systems among respondents, with a tendency towards occasional and non-critical use. This cautious approach could stem from distrust, lack of familiarity, or institutional norms that have not yet fully embraced AI technologies. Among our findings, addressing these barriers through awareness programs, training, and policy changes could potentially enhance the acceptance and integration of AI tools in academic and professional settings.

When asked again how often using GenAI improved their writing style in academic essays, many students once more maintained that they did not use it (Figure 6).

However, they did expect a growth of the use of these technologies, particularly in writing essays and academic papers (Figure 7).

They were resolute in their belief that employing GenAI technologies such as ChatGPT would not compromise the originality of their ideas, with only 22% expressing any concerns regarding its potential impact (Figure 8).

The ambiguity observed in certain responses suggests that students might have been utilizing GenAI tools while being hesitant to acknowledge this usage openly. This is particularly evident in their consistent response to revising essays after employing these technologies, with the majority indicating that they always revise their work (Figure 9).

Upon revisiting the question about the influence of these technologies on their personal style of academic writing, a clear division emerged once more among respondents. Approximately 45% emphasized that their style remained unchanged when using these tools, while 43% reiterated that they did not use them (Figure 10).

Inconsistencies resurfaced when students were queried about the language domains in which GenAI could be beneficial to English learners. Notably, 47% highlighted its potential for vocabulary enhancement, 48% cited improved clarity of expression, 43% mentioned sentence structure refinement, and 23% acknowledged its usefulness in incorporating quotations into academic assignments (Figure 11).

When asked about the optimal method for integrating content from these tools, it was notable that over 25% of students expressed uncertainty regarding its incorporation into their own writing. However, a substantial majority, comprising 70%, asserted that they included content post-revisions while maintaining the integrity of their personal ideas (Figure 12).

Students also expressed concerns about the potential drawbacks of utilizing GenAI in second language acquisition, particularly among students of English Studies. They feared that reliance on these tools might hinder learning and language proficiency development (Figure 13).

The survey findings also indicate that a majority of students trust the information generated by GenAI. This observation appears to be somewhat at odds with their earlier assertion that they revise all content (Figure 14).

In light of the fact that most students trust the information gathered from GenAI tools, it is noteworthy that when questioned again about whether they critically evaluate this information, they affirm that they do. This suggests a nuanced approach wherein students acknowledge both trust in the generated content and their commitment to critically assessing it (Figure 15).

They overwhelmingly assert that they utilize a diverse range of information sources in conjunction with AI tools (Figure 16).

Nonetheless, it is worth noting that 25% of respondents admitted to not ensuring proper attribution of their sources (Figure 17).

They express significant concern about unintentional plagiarism (Figure 18) and fear that excessive reliance on AI could impede their ability to think critically, potentially leading them to accept information without verifying it thoroughly (Figure 19).

Regarding their future employment prospects, students believe that skills acquired through the use of these tools will be applicable to various professional domains (Figure 20).

Finally, when asked if GenAI tools should be banned from university education, 30% students answered that they should, while 44% thought that they could be used as supplementary tools (Figure 21).

The final question in the survey was designed as a checklist to summarize insights and practical recommendations. Students were expected to select all correct answers; however, this did not occur as intended. Students only selected a few assertions as correct, deviating from the intended purpose of the checklist to summarize insights and practical recommendations comprehensively (Figure 22).

• Using GenAI systems like ChatGPT for idea generation and inspiration can help formulate initial thoughts, but the final content should reflect one’s own understanding and analysis. It should be a tool to support research, but never replace personal analysis and synthesis of information.
• One should always provide context when incorporating AI-generated content, connecting the information to your own arguments. Overusing AI may limit adaptability in using a variety of research and writing tools, potentially hindering readiness for diverse academic and professional environments.
• One should always cross-verify AI-generated content critically, evaluating its relevance, accuracy, and appropriateness. Do not overlook the limitations of AI, including the potential for inaccuracies or biases in generated content.
• One should treat AI Chatbots like ChatGPT, as supplementary tools, using them to enhance understanding of a topic, but rely primarily on authoritative sources for substantial content. Overreliance on AI might result in bypassing the process of actively engaging with, comprehending, and evaluating information independently, hindering creative thinking and originality and leading to the loss of personal writing style.
• When using AI-generated text in one’s writing, one should always edit and revise AI-generated content to improve its quality, to seamlessly integrate into one’s own writing style, making clear references to original source texts of information where applicable.
• When using AI-generated text in one’s writing, one should clearly cite and attribute any content in order to maintain academic integrity and allow scholars to trace the information back to its origin. Otherwise, there is a risk of unintentional plagiarism.
• GenAI systems like ChatGPT can guide in problem-solving as well as assist, as a sort of Socratic opponent, in developing discussions with the Chatbot, following the structure of a conversation or debate.
• GenAI systems like ChatGPT can function as personal tutors, giving immediate and personalized feedback based on the information provided.
• GenAI systems like ChatGPT can motivate and facilitate learning by providing a summary of general and specific knowledge on a topic. However, there are still accessibility, equity, and responsible use concerns that need to be addressed.
• GenAI systems like ChatGPT can be used to support language learning. However, for non-native speakers, relying too much on AI for writing in second language acquisition might limit the development of language proficiency.

2.2. Envisioning Future RRI Scenarios

The survey findings offered valuable insights to inform the subsequent foresight methodology and Sci-fi prototyping directed to assess GenAI’s role in the future of Higher Education. The survey highlights students’ attitudes, concerns, and usage patterns regarding GenAI, providing a foundational understanding of their perspectives. It also uncovers students’ apprehensions about unintentional plagiarism, overreliance on AI, and potential hindrances to critical thinking sustaining RRI. Integrating these concerns into Sci-fi prototyping exercises can help envision futures where these issues are addressed, prompting discussions on ethical considerations, regulatory frameworks, and pedagogical approaches. Furthermore, the survey highlights students’ perceptions of the relevance of AI skills in future professional fields. This awareness informs the design of scenarios in Sci-fi prototyping that explore the evolving nature of work and education in AI-driven societies. Overall, by incorporating the insights gleaned from the survey into Sci-fi prototyping exercises, the foresight methodology can more effectively anticipate and shape the future of education in the age of AI, addressing concerns, exploring possibilities, and envisioning strategies for harnessing the potential of AI to enhance learning outcomes.

Additionally, lessons garnered from the survey on GenAI usage among university students significantly contribute to the development of RRI approaches, particularly those emphasizing future-oriented aspects of responsibility. By integrating the survey findings into the foresight methodology, which employs Sci-fi prototyping to assess AI’s future in education, we can align these approaches with the realities of evolving technological landscapes. The survey sheds light on students’ attitudes, concerns, and perceptions regarding AI’s impact on education, thereby informing the anticipation, reflexivity, and inclusion dimensions of responsibility. Moreover, by exploring scenarios where AI’s role in education is envisioned, we can foster anticipatory and deliberative-inclusive responses, reflecting the qualitative aspects of responsibility emphasized in contemporary RRI models [41] (pp. 26–48).

In recent years, RRI approaches present models of responsibility more aligned with the reality we live in. These models focus more on future-oriented aspects of responsibility, rather than demanding accountability for past decisions and actions. They also emphasize qualitative aspects, such as purposes and values, instead of quantitative ones. Responsiveness is identified as one of the four dimensions that configure the concept of responsibility, alongside anticipation, reflexivity, and inclusion. For a techno-scientific innovation to be responsible, it must be capable of generating appropriate responses that are anticipatory, deliberative-inclusive, reflective, and open to change, advocating for broader deliberative forums that involve various social sectors and the necessity to include humanistic and social aspects in decision-making regarding the use of technologies, which is crucial in the AI context. This extends the dimensions of actors and their greater institutional integration, from micro-levels (laboratories) like Complutense Intermedia-Lab to macro-actions (public policies), advocating for broader deliberative forums and the incorporation of humanistic and social approaches in decision-making regarding AI technologies [41] (p. 30). In line with the principles of future-oriented engagement to navigate uncertainty [42], the initiatives undertaken at Complutense Intermedia-Lab have gone beyond conventional approaches by actively involving students in envisioning the future ramifications of GenAI. This approach cultivated a mind-set of anticipatory thinking consistent with RRI principles, facilitating the exploration of AI integration complexities.

The foresight methodology employed a multifaceted approach, incorporating three forward-thinking narratives: two narrative fictions and a graphic novel, along with their cinematic adaptations that delve into potential future scenarios. First, the process facilitator made a keynote presentation of the main trends and breaking points that could affect the university in the next few decades, in order to raise participants’ awareness of the topics that could materialize in the future. Students were then encouraged to individually explore the suggested narratives to understand potential future scenarios influenced by AI, moving on to analysing driving forces and trends, monitoring current events, and anticipating future developments. Subsequently, students engaged in group work and problem-solving activities to develop strategies and responses, gathering intelligence through additional research and formulating solutions through prototyping counter-narratives. Finally, they evaluated the impact of their work to make necessary adjustments. The process comprises the following stages, with data collection throughout the duration of the workshops.

• Explore the narrative: students were asked to read two novels, watch their movie versions, and read also a graphic novel dealing with potential future education scenarios influenced by AI. Special attention was given to RRI-related aspects.
• Explore the dynamics of change: working in groups, students discussed the driving forces and trends that affect these scenarios.
• Monitor current events: students were asked to compare these scenarios with current developments related to AI in education.
• Anticipate future events and develop responses: using Problem-Solving Design Thinking methodologies, students were asked to predict future trends and formulate strategies and responses to address potential challenges.
• Gather intelligence: working individually and in groups, students were encouraged to learn more, collect and analyse information to inform their prototyping.
• Formulate solutions: the groups went on to prototype solutions based on the developed responses.
• Evaluate impact: in sharing their results with other groups, students became involved in the assessment of the outcomes of the process, adjusting goals and objectives as required.

Students were tasked with delving into two novels: Ready Player One (2011) by Ernest Cline [43], and The Diamond Age: Or, A Young Lady’s Illustrated Primer (1995) by Neal Stephenson [44], in addition to viewing their cinematic adaptations. [45,46] They were also asked to explore the graphic novel SciFi.D.I.: Design Intelligence for the Future of Learning developed by a group of experts at Singularity University [47]. A brief overview of these works is provided below.

Ready Player One is set in a dystopian future where virtual reality technology dominates society. The story follows a young protagonist who navigates a virtual universe called the OASIS, where education, social interaction, and entertainment intertwine. The novel explores the tension between the immersive virtual world and the deteriorating real world. While the OASIS offers a superior educational environment, it also represents an escape from the harsh realities that users face outside the virtual realm. The OASIS provides a more accessible form of education, breaking down barriers related to geography, social status, and economic background. Students from different parts of the world can attend the same virtual schools that offer high-quality education with immersive and interactive learning experiences. However, while it democratizes access to education for many, there are still those who cannot afford the necessary equipment or internet access. This reflects ongoing issues related to the digital divide and the disparities in access to technology. There is also the contrast with the declining state of physical schools in the dystopian reality outside the OASIS. Education in the OASIS can be highly personalized. Students can learn at their own pace and explore subjects that interest them in greater depth. The interactive nature of the virtual environment can make learning more engaging and tailored to individual needs. Intelligent Non-Player Characters (NPCs) in the OASIS are powered by advanced AI and serve as tutors, guides, and companions, enhancing the educational experience. The novel also features the character of the IOI’s AI system. It represents the darker side of AI’s potential, used for surveillance, control, and manipulation, highlighting the dangers of society’s dependency on technology, which can lead to vulnerabilities, such as the potential for corporate control and loss of autonomy, as depicted by IOI’s efforts to monopolize the OASIS. The movie adaptation of Ready Player One was released in 2018. Directed by Steven Spielberg, the film diverges in several ways from Ernest Cline’s novel; in particular, the movie emphasizes the contrast between the dystopian reality outside and the OASIS. Spielberg’s adaptation aims to appeal to a broad audience, balancing nostalgia with modern action-adventure. This results in a slightly different thematic focus, prioritizing spectacle and the hero’s journey over the novel’s more detailed exploration of the OASIS as a cultural phenomenon. A graphic novel developed by Cline for Penguin Random House is expected to see the light in 2026.

Neal Stephenson’s 1995 novel The Diamond Age: Or, A Young Lady’s Illustrated Primer explores the use of advanced technology in education and its transformative potential on society. The novel is set in a near-future world and follows the story of Nell, a young girl who comes into possession of an interactive AI book known as the Primer. The book serves as a highly personalized educational tool. It provides customized lessons, stories, and challenges that respond to Nell’s individual progress, reflecting the concept of personalized, interactive, and immersive learning. The Primer not only educates Nell academically but also provides moral and ethical guidance, shaping her character and decision-making. Like Ready Player One, Stephenson’s novel also addresses issues of access and inequality. Nell, who is from a disadvantaged background, gains access to the Primer by chance, as it was originally intended for the daughter of a wealthy family. In addition to the Primer’s automated guidance, human actors (ractors) can interact with the user to enhance the learning experience. This suggests a hybrid model where technology augments but does not completely replace human teachers, emphasizing the importance of human interaction in education. Similar products to Primer are already in the market, known as “Adaptive Learning Technologies” [48].

Finally, SciFi.D.I.: Design Intelligence for the Future of Learning by Singularity University is a forward-thinking graphic novel that explores the intersection of technology and education by presenting visionary scenarios. It seeks to ignite creativity and critical analysis regarding the potential transformation of learning through technological advancements. Singularity University, an academic and innovation hub situated in Silicon Valley, established in 2008 by Peter Diamandis and Ray Kurzweil, is a leading proponent of technological singularity. The institution offers training programs, conferences, and workshops tailored for entrepreneurs and professionals, including the SU Lab start-up incubator. In a two-day workshop held in summer 2019, the university employed foresight methodologies to explore AI’s prospective impact on the education sector over the next fifteen years. The outcome, a speculative graphic narrative set in 2039 titled SciFi.D.I., serves not only as a storytelling medium but also as a methodological tool for envisioning potential futures. With a diverse team comprising over 50 participants, including university faculty, industry experts, and non-profit leaders, the project aimed to provoke reflection regarding the future of education with AI. One of the collaborators is Robert Suarez, who had worked with IDEO (a leading institution in Design Thinking) [49].

In October 2019, members of the OECD Forum community were invited to contribute to the discussion prompted by the graphic novel, posing thought-provoking questions concerning the integration of technology into education, governance considerations, student involvement, curriculum standardization, and the roles of non-profit organizations and foundations [50]. Some of the questions raised included: What technologies would improve the quality of education in the context of lifelong learning? What issues should governance bodies—policymakers, education ministries, communities involved in education—consider? Should students themselves, even the youngest ones like children, have a voice? Should an international organization like the United Nations create a universal, free digital curriculum translated into various languages so that similar content is accessible to people everywhere? Should countries and states continue to establish different curricula? Should a more realistic curriculum, aligned with future labour market needs, be developed? What would be the role of non-profit organizations and foundations?

SciFi.D.I. explores the future of education through the experiences of its protagonists, Carlo, a 7-year-old with a passion for humanities, music, and painting, and his 10-year-old sister, Yabi, who excels in numbers, mathematics, chemistry, and biology. Central to their education are NEPIs (Neoeducational Personal Intelligences), customizable AI companions designed to facilitate personalized learning experiences throughout their lives. These NEPIs cater to individual learning styles, including visual, auditory, kinaesthetic, verbal, logical–mathematical, and emotional preferences, while promoting project-based learning and mentorship in both physical and holographic educational environments. Integrated into a global curriculum network, NEPIs use haptic–holographic technology and nanorobot bodies to mentor students, ensuring continuous learning and skill development. Wearable technology enhances the educational experience by transmitting tactile and proprioceptive feedback to NEPIs, while allowing customization of their personalities to support cognitive growth. Pi, the children’s father, monitors their progress through the NEPIs, emphasizing the collaborative role of families and the education system in setting educational goals. The narrative envisions a flexible education model where students attend class only twice a week, engaging in cross-disciplinary collaboration and virtual reality-enhanced learning experiences.

In contrast to the utopian approach present in SciFi.D.I., dystopian stories often extrapolate current trends to their extreme, allowing us to see the possible negative outcomes, raising awareness about issues that may otherwise be ignored or underestimated. Thus, participants were asked to compare the mentorship situation in SciFi.D.I. to the episode “Arkangel” (Black Mirror Season 4, Episode 2). This episode centres on a mother who implants a device in her daughter’s brain that allows her to monitor and control everything the child sees and experiences. While the primary focus is on parental control and surveillance, the episode raises questions about the impact of invasive technology on a child’s development and education.

3. Results and Discussion

The survey findings reveal several key insights regarding university students’ perceptions and usage patterns of GenAI in academic settings. Despite recognizing its potential benefits, students exhibit reluctance to openly admit frequent use of AI, likely stemming from RRI concerns surrounding academic integrity, negative misconceptions, and fear of being judged as non-responsible users. Lack of clear institutional policies on the acceptable use of AI in academic settings might also be a cause. When utilized, GenAI is primarily employed for tasks such as clarification of texts that students need to read, brainstorming, creativity enhancement, and idea generation, as well as improving writing quality. Interestingly, daily and weekly use of GenAI remains very low across all these tasks, indicating that students do not acknowledge their use, or perhaps that GenAI is still not part of their academic routines. Non-use rates for certain tasks might also imply preference for traditional methods or a lack of recognition of GenAI value and academic acceptance. However, there is an expectation among students for increased future use of GenAI in academic writing. The areas where AI is most impactful (comprehension, idea generation, writing quality) align with tasks that involve cognitive load, although a majority of students (78%) believe that GenAI will not affect the originality of their ideas. This suggests a belief in their ability to leverage AI as a supportive tool without compromising creativity.

Nevertheless, the findings reveal ambiguous situations in students’ use of GenAI technologies. While students are reluctant to admit current use, there is also an expectation of increased future use and a belief that GenAI does not compromise the originality of their ideas. Addressing the ambiguities and concerns through clear guidelines, educational programs, and fostering a culture of transparency could help integrate these GenAI tools more effectively into academic practice, maximizing their benefits while maintaining RRI and academic integrity.

The analysis conducted following the outlined process in Section 2.2 prompted students to discern similarities and disparities in the portrayal of AI’s role in education across three distinct narratives: Ready Player One, The Diamond Age: Or, A Young Lady’s Illustrated Primer and the graphic novel SciFi.D.I. (Design Intelligence for the Future of Learning). In each narrative, AI technologies are harnessed to personalize learning experiences, be it through the NEPIs in SciFi.D.I., the Primer in The Diamond Age, or the virtual reality education system in Ready Player One. These educational tools are designed to provide lifelong learning opportunities, adapting to individuals’ evolving knowledge and skills. Moreover, AI facilitates global connectivity in education by enabling networking and collaboration among students and educators worldwide through virtual platforms. However, the narratives diverge in their storytelling approaches and thematic emphases. Ready Player One is an adventure narrative set in a virtual realm, interweaving pop culture references with themes of escapism and resilience, whereas The Diamond Age combines elements of speculative fiction with social commentary on education and culture. Conversely, SciFi.D.I. uses a graphic novel format to delve into speculative futures and Design Thinking within the realm of education.

Employing foresight methodologies in critically analysing these narratives yielded a set of parameters that shed light on potential future scenarios concerning AI’s integration in Higher Education. These parameters serve as themes, motifs, and symbols within the narratives, encapsulating diverse human motivations, conflicts, and aspirations pertinent to future-oriented analyses. Participants at Complutense Intermedia-Lab identified several such patterns, reflecting societal apprehensions, aspirations, and ethical quandaries surrounding AI’s development and its prospective impact on humanity’s educational landscape.

○

The Mentor: This parameter delves into the concept of AI acting as a mentor or tutor, guiding learners throughout their educational journeys and customizing educational content and methodologies to align with their individual needs and learning preferences.

○

Lifelong Learning: AI technologies are portrayed as facilitating lifelong learning, accompanying individuals throughout different life stages and furnishing continuous learning opportunities to aid in their pursuit of knowledge and skill development.

○

Immersive Learning: The narratives investigate the potential of AI-driven virtual reality simulations as educational tools, providing immersive and interactive learning experiences.

○

Global Connectivity: Across all narratives, AI fosters global connectivity in education, promoting collaboration and enhancing resource accessibility among students and educators worldwide through digital platforms.

The parameters identified in the analysis of the narratives correlate with the findings from the survey on university students’ perceptions and usage of AI in education. Firstly, the Mentor resonates with students’ recognition of AI’s potential to personalize the learning experience, as revealed in the survey. Just as AI mentors in the narratives tailor educational content to individual preferences, students acknowledge AI’s capacity to adapt to their unique learning styles and needs. Similarly, the Lifelong Learning parameter aligns with the survey’s indication that students perceive AI as facilitating continuous learning opportunities. The narratives depict AI technologies accompanying individuals throughout their lives, mirroring students’ expectation of increased future use of AI in academic writing and their belief in AI’s role in lifelong learning. Immersive Learning echoes students’ recognition of AI’s value in enhancing learning experiences. Lastly, just as AI fosters collaboration and resource accessibility among students and educators in the narratives, students anticipate increased AI integration in academic writing and envision AI enabling global connectivity in education settings.

The participants also envisioned unlikely events with the potential for significant impact on the university’s future activities (black swans). These scenarios represent their anxieties or aspirations about possible changes in Higher Education that could drive such transformations. The worst-case scenario threatening education in the future could be the complete domination of AI, leading to a dehumanized, inequitable, and overly controlled educational system. This could lead to a lack of emotional and social development, as students could miss out on the mentorship empathy and the personal connections that only human mentors can provide. Educational resources could become increasingly centralised and controlled by a few powerful tech companies. Wealthier students could have access to superior AI-driven education, while underprivileged groups would be left behind, exacerbating existing inequalities. Economic displacement could also cause social unrest and diminish the societal value placed on the teaching profession. Because AI systems prioritize efficiency as well as standardized and optimized learning paths, this could inadvertently discourage diversity, critical thinking, and intellectual curiosity. Additionally, overreliance on AI could make the education system vulnerable to technical failures, cyber-attacks, and other technological disruptions, jeopardizing the stability of educational systems.

4. Conclusions

The rapid advancement of AI applications in education has surpassed many speculations about the future of learning. To ensure that these technologies are governed responsibly and their impacts align with societal values, institutionalizing prospective methodologies and fostering diverse collaborations with public and private entities is crucial. Prospective methodologies, guided by the principles of Responsible Research and Innovation (RRI), anticipate and manage potential risks, aligning research and innovation with public values and well-being.

The article has shown that by leveraging quantitative and qualitative analyses (surveys + foresight methodologies), we can proactively address future challenges, ethical dilemmas, and societal impacts associated with AI integration in education. Among the findings, the need for clear guidelines on the acceptable use of GenAI tools, along with educational programs to train students on their ethical and effective usage, can mitigate fear and uncertainty among students. It is also necessary to promote a culture of transparency, and open dialogue about GenAI usage can foster a culture of accountability and understanding. Continuous monitoring and evaluation of AI’s impact on student work could ensure that its benefits are maximised, addressing potential pitfalls like over-reliance or diminished creativity. The foresight activities developed at the Intermedia-Lab at Complutense are currently being expanded in the context of “effectuation theory” in the project “PROMISE: Developing Professional Noticing as an Essential Skill for Entrepreneurs”, a consortium of six European Higher Education institutions [51]. Effectuation is a way of thinking and decision-making, based on the idea that entrepreneurs create their future by taking action and making things happen. It is, thus, a foresight methodology.

To sum up, the foresight methodologies explored in this article have offered valuable insights into potential future educational landscapes, provoking a critical and ethical reflection on a wide range of aspects, including students’ autonomy, privacy, equality, and the moral implications of AI integration in education. However, the study acknowledges constraints; for instance, on its reliance on Western sources for the particular actions presented in this article. Other projects undertaken at the Intermedia-Lab have used foresight methodologies to explore non-Western scenarios and RRI perspectives [52]. There are also important limitations in the study because it was carried out in a university context, using only a reduced group of Humanities students. No doubt, this has significantly influenced the outcomes, making them highly segment-specific. Humanities students, with their focus on critical thinking, qualitative analysis, and interpretative skills, may exhibit distinct interactions with and perceptions of GenAI compared to students of technological sciences, who are likely to emphasize technical proficiency, algorithmic comprehension, and practical applications. These factors shape their engagement with and responses to GenAI tools, which is important to consider when interpreting the results.