Business Impact 2024 | Volume 22 cover story: A smart approach

Generative AI is a truly extraordinary technology and is rightly dominating current discourse. To consider the broader potential of AI in education however, we need to examine the technology in its entirety and include the possibilities afforded by AI-driven data analysis and automation. When we zoom out, we see that the ultimate potential of AI is not just to assist with certain tasks, but to perform the full range of tasks of which human minds are capable.

AI is already able to perform every task currently undertaken within our schools – from marketing and admissions to programme delivery and alumni engagement. In other words, it is already possible to create a fully automated business school. This may sound like an astonishing claim, but we need to factor in how platforms such as Coursera already offer a fully automated educational experience.  

While very few people would claim that Coursera offers an equivalent experience to a campus-based MBA, the addition of AI can already improve such automated courses to the point that they can be considered serious competitors to many more traditional, online degree programmes. This capability will inevitably increase as the technology matures.

While every job within our schools can now be performed by AI, this is not a straightforward, like‑for-like automation: AI will perform tasks differently, sometimes better, sometimes worse. That said, just because a job can be automated doesn’t mean it will be – far from it. There are more factors at play than technical feasibility and the degree of automation will vary significantly across our sector.

AI in management education

Each week, those of us in management education encounter at least one new AI start-up and new AI functionalities added to the software we use, plus new pedagogical approaches being proposed. Most of these tools and opportunities can be categorised into five groups, which we've organised here by ease of adoption – from the simplest to the most challenging.

Digital assistants: Most current discussion on AI in education is focused on new tools that enable us to perform our work tasks more efficiently, or to a higher standard. AI tools can assist educators with tasks such as curriculum development, teaching delivery, the provision of feedback and in grading assignments. These digital assistants do not replace educators but rather support them, enabling schools to layer AI on top of existing processes without major structural changes. This category of AI is the most accessible and least disruptive, making it an ideal starting point for many institutions.

AI for analytics: Many schools already use data analytics to some extent. However, any new digital system is unlikely to be adopted unless it provides ready access to the data it generates. The real potential, though, lies in integrating multiple data sources to deliver real-time, actionable insights.

For example, at Imperial College we are working on linking data from 17 different systems that support the student experience. By connecting these data sources, we will be able to understand which elements of a programme contribute most to key outcomes such as student satisfaction, job placement and academic performance. This kind of integrated, analytics-driven approach enables evidence-based course design, more informed decision-making and real-time insights into activities within our schools.

Automation and semi-automation of tasks: Once a school’s digital infrastructure is robust and its data sources are integrated, it becomes possible to automate various tasks. This could include automating responses to student enquiries during the admissions process, sending automated communications to students who may be experiencing difficulties, providing personalised curricula and matching students with job opportunities. As the sophistication of AI increases, so too can the complexity of the tasks that are automated. While this level of automation requires a solid digital foundation, it can lead to significant improvements in both the student learning experience and operational efficiency.

Algorithmic management: At this stage, AI begins to take on a more directive role, akin to the way algorithms manage tasks in sectors such as healthcare or logistics. Algorithmic management involves AI systems allocating tasks to educators or students based on data-driven insights. This approach, while potentially unsettling due to the reduced human agency it may entail, offers the potential for more personalised and efficient experiences.

System-level AI: The final stage is a fully automated, AI‑driven educational system, analogous to ‘lights-out’ factories in manufacturing, where robots completely manage production and maintenance. This could potentially revolutionise education by enabling high-quality learning at scale and at very low cost. This would represent a solution to some of the major challenges in education, such as the need for societies to train many more workers in advanced digital skills. Platforms such as Coursera and innovative models such as that implemented at the 42 network of schools hint at what’s possible: large-scale, low-cost education with minimal human involvement.

It is likely that the adoption of technologies at the start of this spectrum will occur more quickly and perhaps more broadly than those towards the end. However, it is also noteworthy that the value created by AI varies across the spectrum; the more disruptive the AI, the more difficult it is to adopt, but the greater the potential value it can generate.

To read and/or download the full article, please see ⁠Business Impact Issue 4 | 2024 | Volume 22