AI a Challenge to Adult Education

In this blog post I introduce why I consider AI being a grand challenge /wicked problem to adult training and education. Although online learning, augmented reality (AR) and virtual reality (VR) have made significant changes to learning materials and learning environments, in my opinion, there’s nothing as powerful and disruptive as artificial intelligence (AI) once it is fully adopted to education.

AI shaping the education industry and impacting society

AI is a technology that makes use of algorithms when mimicking skills of higher human thinking based on deep learning and reinforcement learning that utilizes big data. AI is making an impact across all economies and is expected to affect the way of working, how people make decisions and forecasts, as well as people’s motivations and attitudes. AI does not only change the way people work, but it forces organizations to change fundamental processes and consider how this new revolutionary technology is embedded in organizations (Holtel, S. 2016).

AI in adult education: Addressing SDGs

The challenge of AI in adult education connects with the Sustainable Development Goal (SDG) number four “Ensure inclusive and equitable quality education and promote lifelong learning opportunities”. More precisely, AI in adult education especially relates to targets that relate to people’s access to affordable and quality education (target 4.3.); technical and vocational skills development for decent jobs and entrepreneurship (target 4.4.); and ensuring learners acquire the knowledge and skills needed to promote sustainable development (target 4.7.). (United Nations). Additionally, I see the wicked problem linking with SDG 8 “Decent work and economic growth”; SDG 10 “Reduced Inequalities” and SDG 9 “Industry, innovation, and infrastructure”. (United Nations). The wicked problem also links to all ESG pillars, mainly social, but also governance and environmental.

What makes AI a wicked problem in adult education?

Every wicked problem is unique and interconnected with other problems. Trying to solve one problem will most likely require solving the previous problem, which your problem might be a symptom of. Due to the complexity of a wicked problem, there’s a risk you might miss something important that happens within the system such as political changes or regulation changes. Due to the complexity of these grand problems, regulatory changes or any individual solutions won’t solve a wicked problem.

The challenge that AI creates in adult education is complex and interconnected with several problems. The individual problems addressing the problem of AI in adult education include ethical dilemmas, inequality, teacher and instructor roles, learning quality and in general the societal impacts of AI in education which we are still very unaware of.

The characteristics of grand problems or wicked problems as often referred to as an overused term, is that the wickedness of those varies according to its dimensions. (Alford, J., & Head, B. W. 2017). Wicked problems differ from “tame problems” with its ambitious and strong conflict settings. There is no agreement or common understanding between different stakeholders on the problem nor on the solution. Therefore, it is impossible to define and solve a wicked problem. (Dentoni, D., Bitzer, V., & Schouten, G., 2018). A tame problem doesn’t mean that the problem wouldn’t be complex as it can be very e.g. technically complex. The characteristics of a tame problem are that it can be described with a clear problem statement. It also has a clear stopping point when the problem has been solved with a solution that is clearly right or wrong. (Holtel, S. 2016). Alford and Head’s (2017, Figure 1) matrix about complex problems presents how increased complexity of problems (y-axis, vertical axis, problem itself) and Increasing difficulty regarding stakeholders/ institutions (x-axis, horizontal axis). The purpose of the matrix is to open the wicked problem for further investigation.

Alford and Head’s (2017: 402-403; Figure 1)

When analyzing the wicked problem of AI in adult education according to the matrix, I notice that it is characteristic to the wicked problem that education includes multiple parties each with some relevant knowledge on embedding AI into education. However, there is no consensus on how to proceed with the challenge, nor on what the problem is. I think that some parties acknowledge AI in adult education as merely an opportunity and are blinded to its problems. AI in adult education offers education institutions the possibility to edit and curate learning material fast and to evaluate learning outcomes, but AI also creates a significant challenge for material accuracy and pedagogy.

I think that AI in adult education meets several conditions of wicked problems presented in Alford and Head article (2017). – Knowability, as AI is still relatively new in education and the problem is impossible to fully understand as there are multiple variations to AI in adult education such as technology (algorithm biases, access, data security), regulations (or those absence), motivations and people’s socioeconomic status. – Knowledge fragmentation, AI developers might be unaware of the specific requirements such as pedagogy or the taught subject. To some extent transparency is missing from AI technologies, making it hard to understand what kind of biases AI might create in education. As AI is changing the majority of job roles, there is a lot of uncertainty about what skills and knowledge is relevant to the workforce and therefore should be taught.

– Interest-differentiation, the interests or values between stakeholders related to education vary. There’s a risk that technology companies prioritize data collection and monetization and policymakers efficiency over ethical use of AI. At the same time, stakeholders from the professional field expect significant productivity improvement through AI and might focus on optimizing it without understanding what challenges AI in their employee training might create.

– Power-distribution, technology companies have significant power over other stakeholders as the owners of technology especially if there are little regulations and little transparency required from the technology companies. Adult education crosses borders through collaboration with different institutions and through international learning programs. It starts to be a politically turbulent problem as AI expands more to the tools and learning methods in the education industry and policies are missing. Businesses are self-monitoring their code of conduct without direct interference from the state or from the civil. According to stakeholder theory those who contribute to a company's wealth can make social or environmental requirements for a business. These stakeholders are seen as beneficiaries. Those are also the stakeholders that influence business decisions. (Steurer, R. 2013). Referring to the wicked problem of AI in adult education, companies developing and deploying AI technologies can currently set their own rules and standards without government oversight. To a certain extent, exploit the opportunity of open markets, somewhat like technology companies did before data and privacy protection legislation. The standards and self-regulation that technology companies are now setting can influence political decision-making and education regulations at later stages, as described in stakeholder theory.

Conclusion

According to the reviewed articles, there are better and worse solutions to wicked problems, but those won’t solve the problem. If a wicked problem has a solution that fixes it, it’s not a wicked problem. Due to the complexity of the challenge of AI in adult education, I define it as a wicked problem.

This blog post is based on a Business and Society course assignment at the Hanken School of Economics. It is the first post of a series of four posts discussing the grand challenge (wicked problem) of AI in education and training.

References

Alford, J., & Head, B. W. (2017). Wicked and less wicked problems: a typology and a contingency framework. Policy and society, 36(3), 397-413. (Accessed online 8th September 2024) Holtel, S. (2016). Artificial Intelligence Creates a Wicked Problem for the Enterprise. Procedia Computer Science 99, 171-180. (Accessed online 9th September 2024)

Dentoni, D., Bitzer, V., & Schouten, G. (2018). Harnessing wicked problems in multi-stakeholder partnerships. Journal of Business Ethics, 150, 333-356. (Accessed online 9th September 2024) Steurer, R. (2013). Disentangling governance: a synoptic view of regulation by government, business and civil society. Policy Sciences, 46(4), 387-410. (Accessed 8h September 2024)

United Nations https://sdgs.un.org/#goal_section (Accessed online 9th September 2024)