We’ve entered, quite suddenly, into an era of pervasive artificial intelligence (AI). We’ve known for a while that it was coming, and we spoke about it in anticipation and trepidation, without clarity of its shape, form or potential.

But when it rather unexpectedly “arrived” — to the public at least — at the end of 2022, we were blindsided by its ease of use and its sheer potential to influence almost all sectors of our lives and economies.

We were immediately faced with our “humanness”, determined to grasp what sets us apart and in what way we could outcompete this new intelligence, however artificial.

Since the release and wide-spread accessibility of large language models (LLMs) such as ChatGPT, Bard, Claude, etc, the way we interact with technology, and our expectations of it, has fundamentally shifted. Whether this is constructive or destructive depends on us and how we respond and adapt to it within the contexts which we operate.

While uptake of the technology in education has thus far been comparatively slow and methodical, businesses and the corporate environment have been swift to adopt AI, using the many technologies and platforms that are increasingly available to significantly improve productivity and simplify their operations. However, such implementation is not without challenge and concern.

Balancing the risks

Apart from the broader existential risks associated with AI, such as the loss of the human connection and the real danger of job displacements, there lie several other concerns that we must urgently, but thoughtfully consider.

Some of these concerns lie in privacy, the lack of AI regulation, misinformation, misinterpretation and manipulation through AI, and, crucially, the biases, discrimination and ethics infringements that presently accompany the technology.

But perhaps most important to consider is the potential of such technologies to either narrow existing inequalities or accelerate them irremediably. For me, this should be our urgent call to action and presents our greatest opportunity.

The Nobel Laureate Joseph Stiglitz recently pointed out that given our current political trajectories and global economic incentives, unregulated AI is likely to worsen inequality

Drawing on historical parallels, technology adoption has always been the driver of flux in equality and access among global societies. Those societies with innovators, developers and early adopters will naturally advance relative to consumers of those same technologies, and AI is no different.

Given that South Africa is, according to the Gini index in 2021 (based largely on income), already the most unequal country on earth, we can reasonably predict the impact of a disruptive technology such as AI on our society, and the inequality gap this is likely to lead to if not addressed through equalising opportunities to access.

Recent disruptors such as Covid-19 and load shedding have already placed the heaviest burden on the most vulnerable, scrubbing off any gains in our transformation and social justice objectives.

So how can we use AI as a vehicle to drive social justice, reduce inequality, and positively transform our society?

AI in higher education

Well, there are the obvious sectors for implementation such as the job market, healthcare, justice and social services, and of course, education. More specifically, it is in the higher education landscape, through our universities, that the most meaningful and long-term impacts of constructive AI integration can be achieved.

Universities in South Africa have all responded to AI, to a lesser or greater extent, through guidelines, seminars, professional and academic development sessions, and through workshops. These have primarily focussed on understanding the technology and its implications, building basic proficiency in its application, and working through policies and guidelines to prevent plagiarism and unethical use by students and academics.

One can imagine the difficulty in ensuring that the myriad of student submissions are AI-free and are authentic student accounts of meeting learning outcomes. Many academics were and continue to be apprehensive, taking a very cautious approach of even banning AI use in their classrooms.

The majority have resorted to the “fight AI with AI” method of relying on AI-based detection tools, which are notoriously untrustworthy, with faulty algorithms that are yet to catch up to the detection level required.

In fact, AI detection tools may be even more nefarious and discriminatory, a bias that will certainly widen inequality for non-native English speakers (of particular concern in South Africa), as this study and this study show.

“Flummoxed’ and ‘perplexed” are descriptions that come to mind to depict the current environment in higher education, which is an uncomfortable position for authoritative scholars.

University academics, as reflective practitioners, are expected to take a deliberate and thoughtful path to AI adoption. Presently, the discourse focuses on pedagogy, exploring many novel and creative ways to integrate AI into learning, teaching and assessment strategies, and there’s been a significant increase in research publications in this space.

Yet, universities are in a unique position to greatly expand AI awareness, literacies and competencies. Rather than only teaching through AI as a pedagogical objective, we have a duty to create equal opportunities for AI access, and must work towards creating a future workforce that are innovators of AI and who can find success in a world that is increasingly shaped by AI in almost all sectors of the economy, from agriculture, to finance, healthcare, design, and everything in between.

AI literacy

This will also lead to new career paths and opportunities for the AI-literate, who will have a distinct advantage over those who will be deemed “AI-illiterate”.

And with a wider understanding of the implications of AI across all spheres of work and society, we, as the Global South, will then be in a position to innovate and command the direction we want AI to take, aligned to our needs and shaped for our purposes, rather than being passive recipients.  

For this to happen, however, an overhaul of our curricula is a necessary condition. We need to integrate AI literacies and AI practice, through AI pedagogies, into all curricula, across disciplines, and further into our delivery and assessment strategies.

We must realise a critical shift in our approach to higher education, while retaining the fundamental ideals of student support, personal values, attention to needs, and holistic wellness of academics and students.

Indeed, university support structures can also greatly benefit from AI integration and the big data that can be retrieved from it for greater efficiencies in operations and management, and for a clearer understanding of diverse student needs.  

This is not to say that AI isn’t already in many of our curricula. However, these are inconsistent and concentrated in certain disciplines, like the applied sciences and engineering, that are context-bound, whereas AI is in fact socially bound.

Hence, just as AI is pervasive in society, AI literacy should pervade across our curricula — humanities, social sciences, law, philosophy – every discipline has a role to play and a contribution to make, all guided by constructive and meaningful university policies.

To achieve this, the first condition is for all universities in South Africa to work together on a re-curriculation project that spans all disciplines. We’ve internalised the mantra of “no student left behind”, but need to move towards “no institution left behind”. We must reduce the inequality among institutions and build AI capacity for all, by creating national think tanks to collaboratively drive this.

I can think of structures such as the Higher Education Learning and Teaching Association of South Africa (Heltasa) and the Teaching Advancement at Universities (TAU), the latter with a specific social justice objective — to be custodians of this project, given that they are represented by members of all institutions, and have the capacity to coordinate such an initiative.

The Council on Higher Education (CHE) has oversight and ensures quality of design and delivery, and measures the impact of such a ‘re-curriculation’ initiative. Without central coordination, however, we risk widening inequality among institutions, and hence huge disparities among graduates in just a few years.

With revised curricula that are future-fit, we can then move to resource allocation that sees investments into super-computers in each region, which service multiple institutions. This will further encourage collaborations and can also serve industry needs.

Implementing a dual education model, for example, would create learning spaces and campuses that are connected to the relevant industries so that learning and teaching occurs within the context of work. This will in turn raise the value of our universities, with the concomitant enhancement of graduate skills and attributes that are work-ready.

The prevalence of AI has presented a truly unique opportunity for universities to effectively re-imagine their strategies and position in a fast-transforming world. As custodians of the knowledge economy, our institutions have no choice but to lead this transformation, and intentionally chart the direction it takes, through reflection and meaningfulness.

It’s our best chance yet for innovation and social justice. If we miss it now, we’ll be forever consumers. DM