Accessibility Tools
Can universities move beyond present discussions of AI’s disruptive impact on higher education and embrace this (forced) opportunity to rethink our role in the society of the future? For Manfred Krafczyk, this process may be cumbersome and even painful, but it will eventually allow an unprecedently deep and wide reshaping of our institutions and scientific disciplines.
While the fundamentals of artificial intelligence (AI) are rooted in developments in Computational Sciences from the middle of the twentieth century, recent technical advances in AI have exponentially broadened its range of applications to essentially all scientific disciplines. While the underlying technology expands at a huge rate, at the same time almost all scientific disciplines are faced with the opportunities and challenges of AI. So, how can they recalibrate their approaches to derive new knowledge without sacrificing their discipline-specific principles of quality control?
Adapting to AI can only be accomplished from within each discipline, approaches cannot simply be copied from others. Indeed, the new challenge is to adapt AI technologies to the specific requirements of all scientific disciplines. This challenge is unique for each discipline and not a primarily technological one. The potential beyond mere efficiency gains could be a deep, qualitative change to individual disciplines and the way that they operate.
It has long been suggested that the mere teaching of discipline-specific facts should only be considered a minor part of university education. Instead, we focus on imparting competence as the fundamental requirement for future professionals who are expected to extend their knowledge by self-organised lifelong learning. With AI already at hand in undergraduate education, lecturers now tend to expect that their students are capable of evaluating the quality of AI-generated results while still in the process of internalising the fundamental principles of their respective scientific discipline. This challenge requires a careful analysis of present curricula as well as how they can be sustainably taught utilising AI.
Before the advent of modern information technology, the distinction between knowledge and competence in higher education was rarely a subject of intense discussion. Applying both principles and discipline-specific knowledge to previously unknown contexts and problems is an important criterion of scientific expertise. Therefore, the expert of the past was expected to have internalised both principles and knowledge/data/facts so that they could create new solutions utilising internal information. However, in recent decades, the externalisation of factual knowledge to the internet has arguably diminished the ability of domain-specific problem-solving.
The unguided use of AI tools has the potential to accelerate this undesirable turn. In particular, undergraduate students may assume that it suffices to delegate their know-how to external data-based tools. This would deprive them of their ability to solve problems in real time, utilising their own knowledge and competence. Maybe we should focus on the goal of becoming problem-solvers who occasionally use advanced tools instead of a priori delegating problem-solving to these digital tools as a default approach.
AI tools have already advanced to a level where they are arguably better at certain tasks than any human. Yet, this does not mean that we should simply give up on certain skills. Instead, we should use these extended capabilities to explore new levels of complexity and deeper levels of the underlying fundamentals in our respective disciplines.
While digitalisation has enriched essentially all aspects of higher education, it has also increased our dependence on digital tools, the externalisation of knowledge and data and the influence of the few companies that dominate the field. AI technologies have the potential to either increase or reduce our level of sovereignty depending on how we, as higher education institutions and society at large, decide to use them. If we use AI tools without reflection, we will see the diffusion of biased, incomplete, and incorrect information into our problem-solving approaches. This will directly lead to a loss of sovereignty, and not only in higher education institutions.
There is a reason that a major part of the investment of AI companies goes into quality control of their tools. We must also invest in protecting our institutions by strengthening AI competencies for (especially non-STEM) students, lecturers and researchers, as well as administrative staff, to become AI competent in order to keep, if not increase, whatever digital sovereignty is still left.
AI is not only here to stay, but is likely to substantially evolve in the next decade. That being the case, it is important to reflect upon higher education’s fundamental role in the society of the future.
The worst-case scenario might include that universities (or at least public universities) will have to justify their costs in the presence of seemingly ubiquitously available know-how from AI. Indeed, tech companies already argue that they are more efficient at educating their staff. Not only for this existential reason, we should intensify our efforts to adapt higher education and how we conduct both fundamental and applied research, while integrating and shaping AI according to the highest ethical and scientific standards.
