Across Europe, there is significant debate about the need for Education to address current societal challenges. New employment structures and the need for an adaptable workforce; a changing information and knowledge environment; increasing mobility of people across the globe; concerns about rising inequality; and demographic change are just some the issues that many hope Educational systems can address. This intensification of debate is apparent across academia (Autor and Salomons, 2018), policy (e.g. OECD, 2017), industry (e.g. PwC, 2018) and the third sector (e.g. Schneider and Bakhshi, 2017); encompassing calls for “our education systems (…) to achieve at levels that none have managed to date” (Luckin et al., 2016: 12).

One of the interesting themes within these debates is the way that new technologies, in particular AI, is viewed as both a driving force of current societal challenges (e.g. deskilling and automation of the workforce) yet is also viewed as a potential solution to help Education meet the new needs of society: where AI can be ‘harnessed’ as a tool to facilitate learning opportunities, process and outcomes across many settings and contexts of Education.

Utilising a mixed method approach, and drawing on existing critical and philosophical work on debates about the relationship between Education and Society (e.g. Biesta, 2013; Apple 2012), this talk will examine the ways that AI and Learning are currently being conceptualised by different policy, commercial and academic actors. Focusing in particular on discussions and practices around lifelong learning, we highlight the current limitations of this fragmented discource and propose an alternative direction of travel that focuses on designing AI within lifelong learning contexts to support social justice.

The paper will present data from a 2 year study that utilized three approaches to exploring this topic. The three methods are: A systematic analysis of existing literature on this topic: we reviewed scholarly databases including Google Scholar Scopus and Web of Science using 20 appropriate boolean search targets then merged all results into a single BibTeX database using JabRef. The data was cleaned, removed, and missing data obtained using web data extraction as part of this process. We then visualized the citation network and keyword co-occurrences using VOSviewer, conducted topic modelling using latent Dirichlet allocation and non-negative matrix factorization, navigating and extending each of these steps with qualitative coding of a small proportion of academic outputs. Interviews with key AI and Education experts: we purposively identified 10 academics, 10 policy actors and 10 commercial stakeholders to interview. The interviews were carried out face to face and skype. Each interview took around 45 minutes and were focused on the how AI was currently being used in Education, key challenges and opportunities, the (dis)connections between work across commercial, academic and policy domains and likely future directions. All the interviews were audio recorded and transcribed. The data was analysed using multiple rounds of qualitative coding to refine the themes, visualisation of the data, and testing of alternative explanations (Dey, 1991; Miles and Huberman, 1994). Network analysis: we built an open source Network Discovery Tool to semi-automated the identification of companies, individuals, products, and other activities appearing in news, social media, and websites related to AI and lifelong learning. Based on co-occurrence of extracted entities, we then modeled collaboration networks to be further analyzed and explored using a graph database system. Using this system, we explored the range of actors beyond academia that influence and shape discussions of AI and lifelong learning provision, direction and focus; and the different configurations of actors engaged in the development of AI applications to support and facilitate personalised, effective and scalable solutions for lifelong learning.

In this paper we highlight how lifelong learning is routinely defined as an act of economic self-interest, and the ways in which the AI and lifelong learning landscape is highly fragmented with current debates dominated by economic and commercial interests. In the talk we will propose a more nuanced view by synthesising learning theory, the sociology of educational technology and the affordances of AI. We argue that future work should be more strongly directed towards questions of social justice. While important to prosperity, a preoccupation with the economic and labour market returns of lifelong learning risks marginalising its social, personal, and democratic benefits to individuals and wider society. A more expansive conceptualisation of lifelong learning is essential for the overall ‘health’ of a society (Albeit, 2009; Biesta, 2006; Field, 2006); and crucially ensures the design of lifelong learning strategies to support alternative desirable futures for us all (Facer, 2011; Guile and Livingstone, 2012; Levitas, 2013, PwC, 2018). In an era characterised by AI, where there are many unknowns about how AI and humans could and will interact, multiple futures need to be accommodated in all our designs. More conversations amongst key stakeholders are required about the kind of society we value, and how the shape, nature and support for lifelong learning can be planned with these alternative futures in mind (Apple, 2012).

Apple, M. (2012) Can Education Change Society? Chicago: Chicago Press. Accenture (2016) Artificial Intelligence is the Future of Economic Growth. Accenture (NYSE: CAN) Available at, https://www.accenture.com/futureofAI. Alheit, P. (2009) Biographical learning within the new lifelong learning discourse. In Illeris, K (Ed) Contemporary theories of learning. London: Routledge. Alheit, P. and Dausien B. (2002) The ‘double face’ of lifelong learning: Two analytical perspectives on a ‘silent revolution.’ Studies in the Education of Adults, 34(1): 3-22. Autor, D. and Salomons, A. (2018) Is automation labor-displacing? Productivity growth, employment, and the labor share. Brookings Papers on Economic Activity. Bagnall, R. (2000) Lifelong learning and the limitations of economic determinism. International Journal of Lifelong Education, 19(1): 20–35. Biesta, G. (2013). Responsive or responsible? Democratic education for the global networked society. Policy Futures in Education, 11(6), 733-744. EU Commission (2018) ‘EU Digital Education Action Plan’. [Online] Available from: https://ec.europa.eu/education/sites/education/files/digital-education-action-plan.pdf. OECD (2017) Financial Incentives for Steering Education and Training. Getting Skills Right, OECD Publishing. Paris. Available at, O'Neil, C. (2016) Weapons of math destruction: How big data increases inequality and threatens democracy. Broadway Books. PwC (2018) Workforce of the future. The competing forces shaping 2030. PwC. Available at, https://www.pwc.com/gx/en/services/people-organisation/publications/workforce-of-the-future.html. Pea, R. (1998) Practices of distributed intelligence and designs for education, in: G. Salomon (Ed.) Distributed cognition (Cambridge, Cambridge University Press) Robins, K., & Webster, F. (1989). The technical fix: Education, computers and industry. London: Macmillan Russell, S. and Norvig, P. (2009). Artificial intelligence: A modern approach. Prentice hall Schwab, K. (2016) The Fourth Industrial Revolution. World Economic Forum. Sfard, A. (1998). On two metaphors for learning and the dangers of choosing just one. Educational researcher, 27(2), 4-13.