I gave a talk at the Royal Society’s 2018 annual diversity conference. This is the text of my contribution.
Thank you to all the organisers of this year’s annual conference for this very special opportunity to speak to you all today.
My aim today is to spark a discussion. And I want to do this by exploring two readings of this question with you. An optimistic reading, centred on our role in restoring self-confidence to under-represented groups. And a second, more critical reading, that casts this question as one that detracts us from asking who is responsibility for change.
As you think of this question, consider the journey that is taken by the under-represented groups we might have in mind. They, like me, may come from the other end of the world, from the southern tip of Africa, from a town shaped by segregation, and an education system that could never inspire, yet somehow, come to hold the doctoral degree from our ancient university in Cambridge. They might, like me, be born and live in a time where to be coloured is to accept overt and covert discrimination and disadvantage, yet come to add new thought to their field of science. They might, like me, have to overcome shame and confusion and rejection, to learn to accept themselves and embrace their queerness, and somehow, come to use that to strengthen their institutional culture. Journey’s like mine are our struggle credentials. This room is filled with struggle credentials. Struggle credentials play too much of a role in our present. The sign of the success of our work will mean that over time, these credentials become harder to attain.
The attainment of the credentials of struggle emphasises an important transformation that occurs along this journey. These journeys emphasise the creation, or better, restoration of self-confidence and self-ownership. When the question of decolonisation of science and of our thinking is raised, it is restoration that is meant. So, this becomes the first part of our effort: the restoration of self-confidence and self-ownership. And to instil this uniformly across people.
If I think about today’s question in my own area of artificial intelligence, then I often express it in a different way. The AI that we believe to be global is not. It is localised in individuals, institutions and countries. How do we make global AI truly global? And of course, we can replace AI with any other area of science, or endeavour.
About two-and-a-half years ago, we created a new organisation called the Deep Learning Indaba, as one attempt to answer this question. The Deep Learning Indaba is an independent grassroots movement whose mission is to strengthen African machine learning. And we do this by building communities, creating leadership, and recognising excellence across our African continent. This year alone we solidified the confidence of almost 550 of the brightest African talent in AI at our annual gathering. These future leaders have confidently entered the conversation around AI, their skills, and their contributions to their societies.
This triumvirate of community-leadership-recognition is what I consider as the basis of a self-supporting ecosystem that builds itself, puts itself forward, and claims its ownership in the field of artificial intelligence. As we hoped, this philosophy is now being replicated. Communities are being built and ownership is being taken globally: earlier this year in Eastern Europe, next June in South-East Asia, next November in South America, and hopefully soon in South Asia. Global AI is now taking its transformative steps to becoming truly global.
One of the lessons here is on the importance of grassroots organisations and their collective role in putting forward and supporting diverse groups. The grassroots are those groups within our institutions, like the LGBT employee resource group I lead within DeepMind, and those outside movements, like the Deep Learning Indaba. One of the clear strategies we have available to us is to consider the support we offer to grassroots groups: through supporting their creation, through our funding, sharing our leadership expertise, with our time and networks, and in all the simple encouragements that people need to continue the work of true transformation.
My second reading of this question is a bit different. This question creates discomfort. Perhaps that is its intention.
This discomfort develops because it reveals several other uncomfortable questions: Who exactly are the under-represented groups? What systems of power are at play and pre-supposed by this question? Why can’t these groups put themselves forward already? What does forward mean: to what, to where, or before whom?
For me, the problem with our original question is that it is unquestioning in its outcome. It is the under-represented groups that must eventually be put forward. It is they who must change. It is they who become equal to the groups that are already represented. It is problematic because the value-system of we that might find ourselves currently represented remains intact and sustained. It releases the Represented from responsibility.
The questions we ask, must instead be re-centred: on ourselves. Each of us must assume a personal responsibility for transformation; and the questions we ask must prevent us from seeking release from this responsibility.
Some alternative questions: How diverse was the team of collaborators in our last few projects or papers? Do the events we speak at have the representation we expect or demand? When we organise events, how do we ensure inclusivity and reflect the real forms of excellence that exists. We should all keep a log of our activities to measure and inform our decisions. What are the ways in which we question our own settled assumptions of people and systems?
Simply asking these questions provides us with a personal strategy for change. And so a key principle is revealed.
The work of transformation lives in the mundane. Our personal responsibility is taken in the everyday, in the simple decisions, in our small actions and reactions. When transformation lives in the mundane, it becomes magnified in the larger structural and organisational systems that we influence.
But transformation living in the mundane means that our effort will not be without sacrifice, and not without struggle. We who have the privilege of representation will have to give up our place. We will have to give up our recognition. We will have to, at times, not be recognised at all. We will need to divert power away from ourselves towards others. Our human nature means that this price is amongst the highest we could ever be asked to pay. But it is ours to pay, now. And a debt to the future we can erase.
My aim has been to put forward two principles for our debate. Firstly, the restoration of self-ownership and the importance of the grassroots. And secondly, on personal responsibilities and the price of transformation.
For everything we will be asked to pay, there is a reward. And that is a source of sustainable and lasting joy. I think we experience a small part of this joy through the celebration of progress that has brought us here today. And I hope for, work towards, and expect, much more joy to come. Again, thank you for this opportunity, and your attention.
The material for this talk was created using some of my writing in two other essays on The Price of Transformation and Decolonising Artificial Intelligence.–At the Royal Society, considering the mundane.
Related