A New Consciousness of Inclusion in Machine Learning

On LGBT Freedoms and our Support for Machine Learning in Africa

This is an exploration of my thinking and my personal views. Read in 5 minutes · 1147 words ·

Soon, in two neighbouring countries in Africa, two large machine learning events will take place: one in Nairobi in Kenya, and another in Addis Ababa in Ethiopia. The choice of these host countries has fomented concerns throughout our machine learning community: how can we as a community committed to inclusion in every form consider hosting our conferences in countries like these that are far from inclusive? It is specifically the criminalisation of homosexuality and the social and legal status of LGBTQ+ people in these two countries that is our cause for concern. A politics of location, and an ethics of inclusion is growing healthily within our machine learning community. This puts us in a special moment in the history of our field.

I find myself with the great honour of a place in the leadership of these historical events, and tasked with their success. It is the International Conference on Learning Representations (ICLR) that will be held in Ethiopia in April 2020, and the Deep Learning Indaba that will be held in Kenya in August 2019. The concerns that are raised do not prevent me from being a passionate supporter of the change that these moves represent. I am of course a proud South African, a committed pan-African, one of the Indaba’s founders, and the senior programme chair for ICLR2020, so my support is not surprising. But I too am an out and proud gay machine learning scientist. The concerns of safety that underlie these questions are more than hypothetical for me. Yet, rather than shy away from decisions and locations like these, we are now given the opportunity to build a new consciousness of inclusion that embraces these changes and the complexities that they come with. While there is genuine reason for concern, there is also so much to gain.

At the outset, we must hold to the principle of safety first. Supporting these locations means that no person should attend in person if they cannot be assuaged of their safety. A lack of legal guarantee means that for many in our LGBTQ+ community, this may not be possible. But if possible, there are ways in which we can make this assessment: by looking to the specific security conditions that are put in place, the on-the-ground behaviour and record of enforcement and violence, through the guidance given by within-country activist groups, and by learning about the lived experiences of people in these countries. And this never means pretending to be something we are not.

As the writer Frankie Chike Edozien says in his book The Lives of Great Men, “Today, gay folk who live in African cities have very full lives---challenging, yes, but full nonetheless”. Part of our work must be to engage with the realities of gay life in Kenya, and in Ethiopia, and perhaps of the continent more generally. How else will we see that current political complexities do leave hope for change: with the former Nigerian president, Goodluck Jonathan, once his term ended, stating that “with the clear knowledge that the issue of sexual orientation is still evolving, the nation may at the appropriate time revisit the laws”; or that former Mozambican president Joaquim Chissano spoke against the Nigerain bill at its passing with a clear call for gay rights across Africa; or that Botswana's former president Festus Mogae has done the same. And while we are disappointed by Kenya’s recent court ruling, it signifies a coming change---a change that took place just over 50 years ago in the UK itself, just about the time of Kenya’s independence from colonial rule.

Hosting and sponsoring our conferences in countries like Kenya or Ethiopia does not mean we endorse their governments or their politics. And while we must operate within the legal frameworks we find, the question we must ask is, who benefits from our absence and relegation of these places? It is not the young entrepreneurial developer communities in these countries who continue to be excluded; it is not the universities who are rebuilding themselves and who will now fail to be confronted with new ways of working and organising; and it is not us, who would instead be denied an exploration of the ways of living and loving in countries other than our own. The concern and activism we espouse online must be transformed into action in the offline world, and this can certainly be achieved by organising in countries like Kenya and Ethiopia. A passive sanctions or twitter protest alone will not contribute to change. We need to develop a double vision and a double politics that allows us to advance the cause of an inclusive future, while working with the challenges of the present.

For if we are truly committed as a community to the ongoing mission that AI must benefit everyone, then it becomes imperative that we engage in the intercultural dialogue that this demands. How else will we understand the reality of multilingual Ethiopia, and the tension that Amharic, its lingua franca, creates as a uniter and divider of the country. How else will we understand the social reaction to the negative impacts of poorly thought-out technology in Kenya, which have led to botched elections and political upset. Simply being in countries outside the usual means that we begin to put in place safeguards that will help us avoid being part of a technological colonialism, in which our countries and companies and organisations extract value and money, while never engaging in any meaningful way in the transformation and success of the countries we currently shun; taking resources and talent and money and leaving nothing behind.

Ultimately, the support we provide and its form is up to our individual consciences. Nothing is binary, and our choice is received without judgement. LGBTQ rights are only one of the areas of concern in Kenya and Ethiopia, and elsewhere globally---women’s rights, free speech, and religious and tribal tensions remain. And neither are we free from ethical quandary ourselves, with conflict minerals in our phones, through a reliance on clothing produced under suspect labour conditions, and on jet fuel and petrol and their questionable sources that sustains our best conference lives.

My hope is that we will always continue to experiment with the ways in which we organise and support our global machine learning community. And as we do the work of change in the real world, we will build a new consciousness of inclusion that will infuse our science, our products and our impact with the vision we need to support the brighter future to which we all work.

If you are interested in reading more of my writing related to transformation and diversity, read or watch a talk on Racialised Lives and the Life Beyond, or on  How to Support Under-represented Groups to put Themselves Forward, or two other essays on the Price of Transformation and Decolonising Artificial Intelligence.

-- Considering choices

One thought on “A New Consciousness of Inclusion in Machine Learning

Leave a Reply

Your email address will not be published. Required fields are marked *