Between Cyberutopia and Cyberphobia : the Humanities on the African continent in the Era of Machine Learning

Thursday, 7 March, 2019 - 09:00
Add to Calendar

Draft Call for Proposals

This workshop takes its focus from the upheaval in popular and scholarly understandings of the intellectual (and political) prospects of the networked planet. A decade ago advocates and precocious users were celebrating the levelling, democratic and disruptive possibilities of the Internet, and of social media platforms in particular.  Today an elaborated loathing of these technologies and their political effects -- succinctly captured by the Channel 4 series Black Mirror -- has become common and politically compelling.  Popular and scholarly disillusionment with the promises of the network society is now close to self-evident.  Much more difficult to assess is the critical and political power of the dystopian critique of cybernetics (of which Black Mirror is only the most recent compelling example) that emerges from the humanist tradition.  In this workshop we propose an assessment of these two movements, and their mutual engagement, in the special circumstances of the African university.

The combination of ubiquitous social media, feedback-centred devices and the sorting and predictive techniques of machine-learning, now released from the old constraints on data-processing, seems to present an existential danger to the long-established habits of disciplinary enquiry in the humanities. These carefully engineered features of the global network affect young and old alike. An explosion of source materials -- to focus only on the most obvious problem --  has been combined with a technological order of continuous distraction, where user-produced data (much of it in text form) has become the key profit-driver for the wealthiest firms in the world. In this new global economy, and in the simplest formulation, it is the absence of uninterrupted time that makes the spontaneous development of the post-Renaissance ethic of self-directed reading increasingly untenable.  It is this predicament that drives the deep and persuasive pessimism about the production, consumption and meanings of digital content, and devices, that now dominates the humanities.

The South African universities confront this predicament, and the humanist’s problem of encouraging contemplative, self-directed reading, constrained in unusual ways. The universities here were formed by an intellectual history that sought to limit the power of books -- especially those written by black authors, by limited public investments in libraries and in popular literacy, and by the extreme forms of the inequalities that structure reading ability everywhere. Commercially produced books are today objects of elite conspicuous consumption, and they are far out of reach of all but the wealthiest undergraduate students. After twenty years of compensatory policy and spending, South African schools remain powerfully shaped by the Bantu Education curriculum and racist resource allocations that affect book cultures directly.

The digital transformation of reading that has taken place over the last twenty years – the development of massive free (and some pirated) on-line repositories like Jstor, the Internet Archive, Sci-Hub, and the Genesis Library – presents a magnificent opportunity.  All South African students now have potential access to a vast digital library of books, journal articles, films and photographs. And with the added popular momentum of (mostly) text-based forms of social media, South Africans, really for the first time, can begin to chart out complex, peer-nurtured journeys through the vast corpus of the global humanities. The digital upheaval presents the first real opportunity to foster a democratic humanism in South Africa, and many other countries on this continent.

Yet the boundless wealth of textual information also represents a new kind of obstacle to the development of exploratory reading. The constraining (and beguiling) space of the old physical library has itself been demolished by the publication on the web of the contents of the basements of thousands of libraries. The digital equivalent of browsing the shelves under these circumstances can be an extreme form of the rabbit-holing that devoured Alice. Similarly, and despite the enormous scope of the bibliographic collections that are available on-line, there are clear and discernible voids in these collections that reflect what African and feminist scholars have long described as a project of invisibilization. South African students will search in vain for important books by Mafeje, Magubane, Mangcu, Meer or Nolutshungu in the pirate repositories, nor will they find the old, foundational works by Fuze, Rubusana, Molema or even Plaatje in the Internet Archive. The very scope and depth of the on-line repositories can have the insidious and powerful effect of reinforcing the segregated structures of the 20th century academy.  While it is important to be explicit about the intellectual and political effects of these problems of visibility, we are also interested in developing alternatives and solutions.

To this end, in this workshop we will focus on the intellectual and philosophical claims of machine learning as both a subject of humanist enquiry and as the key set of skills and technologies shaping the workings of the network society. Precisely because the field is concerned with the engineering of consciousness the history and contemporary debates in machine learning are philosophically sophisticated, richly informed by the old core problems of humanism, and many of the key technological claims are evaluated, within the engineering disciplines, using the old philosophical arguments.  The most influential and persuasive scholarly accounts of the current state and future prospects of artificial intelligence are those that take this history of philosophical and institutional conflict as a guide to its future. Likewise it is the researchers working in this field -- especially the youngest researchers -- who are most alert to the potential for the new technologies to renew and entrench the oldest structures of inequality, and preoccupied with developing automated remedies.

The combination of machine learning and the ubiquitous digital network also presents a set of remarkable opportunities for young people on the African continent.  There are now many programmes offering mathematically-inclined African students training and exposure in the core skills of machine learning and artificial intelligence outside of the university. The African Institute for Mathematical Sciences (AIMS) draws on the enthusiasm of global experts to support the activities of six schools -- in Cameroon, Ghana, Rwanda, Senegal, Tanzania and SA --  that offer free training in the mathematics and computer science of data science and, increasingly, machine learning. Half a dozen similar and loosely related projects, offering the same kinds of exposure and connections are now operating on the continent, including the Deep Learning Indaba, which is supported by young South African researchers at Google’s Deep Mind AI lab in London. Perhaps the most startlingly elitest of these projects, leveraging the demography of talent on the continent, is the Next Einstein Forum, which draws on the AIMS network to anoint a handful of young researchers from across the continent each year.  This creates, in science and mathematics research, an economy of talent similar to what has long been true in football and music.

It is important to notice that these open and free networks and training workshops are matched to a broad suite of resources that are freely accessible to anyone interested in pursuing ML as a field of expertise. In stark contrast with the other main areas of science and engineering research (which have long been bitterly under-resourced, stripped of laboratory tools and simply unavailable on most of this continent) the tools of ML are all freely available on-line. The most powerful computational platforms -- the same platforms used by researchers at MIT and Cambridge -- are available, for free, to African researchers with access to the Internet.  This is something like a revolution -- the material opposite of what has been true of scientific research for at least a century. The open-sourcing of data for developmental ends (D4D) sometimes supports the public exposure of records that would not be permitted in more carefully regulated societies, and sets in place the conditions for African countries to act as laboratories for global machine learning experiments that would not be possible elsewhere. Whether this distributed and open-sourced network will strengthen the universities and the intellectual and material priorities of the humanities (which AI is doing in part in the elite universities of the West) remains a wide open question.  

There is also, already, a long list of the well-articulated political and intellectual dangers that follow from the development of generalised tools of machine learning. The most obvious derive from the dense, hidden and ingrained structures of racism. There are also problems of bias that result from the absence of high-quality training datasets -- for example of African names or facial images or words. A third obvious risk is that AI will exaggerate the already existing brutal deficits of infrastructure – of high-speed network connections, reliable power supplies, data processing centres and, especially, of human expertise. Many experts worry that the growing power of the centres of artificial intelligence in the United States and China – and the global monopoly power of a small number of firms secured by AI -- will produce a new era of data-driven extraversion and dependency that will remove the decisive philosophical and political deliberations from the continent. And there is the intractable problem of formal work itself in the face of a global movement to automation. A less obvious question for the humanities on the African continent is whether these new tools will support insidious and powerful infrastructures of social ordering.  Companies that specialise in the technologies of surveillance and social scoring that the Chinese state is fostering have already found easy accommodation in the African countries -- including Ethiopia, Tanzania, Uganda and Rwanda -- that share a common vision of bureaucratic control and surveillance and weak privacy laws.

After several decades of naive optimism the humanities have reached a moment of bad-tempered critical reflection that offers productive insights into the strengths of the disciplines, the priorities for the future and opportunities that may still be realised in the networked society.  These questions have particular urgency on the African continent where the weakness of the humanities, the limits of regulatory constraint, the offshoring of data-processing, and elites’, donors’ and states’ enthusiasm for automated tools of surveillance and social ordering suggest the possibility that the networked dystopia that is much worried about in the rich countries may first take form here.   

In this workshop we invite proposals for papers that consider the following or related problems:

  1. Does the universal distribution of attention-mining social media represent an inescapable existential danger to the (often unconscious) intellectual habits on which the humanities have been established?

  2. Does the development of ubiquitous and automated scoring -- of either the Chinese or American kind -- subvert the normative and political ambitions of the core disciplines of the humanities -- of philosophy, political studies, religion, literature?

  3. Will the universally distributed network, and the open-sourcing of the tools and platforms of machine learning, strengthen or weaken African universities and research, and the humanities disciplines within them?

  4. What can we learn from the intellectual history and philosophical debates within the fields of artificial intelligence about the prospects of these technologies and their relationships with the humanities?

  5. What opportunities and remedies are available for those who seek to disrupt the ordering and extractive logics at work on the network?  Can we use the same -- or similar -- technologies to achieve contradictory ends?