Opening remarks delivered on Day 3 of the 100 Brilliant Women in AI Ethics Summit on December 5th 2020. Find out more about Women in AI Ethics TM here.
Hello and welcome everybody to today’s sessions at the 100 Brilliant Women in AI ethics Summit.
My name is Sinit Zeru and I’m deeply honoured to be a member of the advisory panel and to have the opportunity to speak to you.
When Mia reached out to me to participate in this Summit – I had two thoughts:
First: wow, what an honour and an amazing opportunity to learn from officially Brilliant women.
My second thought was: Me? I’m not very techy. I work with techy people to think about policy but I’m just learning about AI ethics myself – do I belong here?
Fortunately, Mia reassured me there was enough tech and AI ethics expertise in the room. In the spirit of this Summit. What she wanted was:
To expand the types of voices in the debate
To enable more diverse thinking on the tough ethical issues that AI provokes, and
To make space for the people who are impacted by these issues and yet woefully under-represented. For this year’s event, I know there has been special effort to include voices from under-represented regions like Sub-Sharan Africa (SSA) – where I’m from and where I live and work.
It’s hard to say both accurate and meaningful things about Africa because of its incredible diversity. There is also the fear of feeding the misperception that it’s a more homogenous place than it really is.
It’s a continent of 54 nation states. To narrow that just slightly, I’ll focus on Sub-Saharan Africa because that’s the region I’m more familiar with. But even then, that’s still 46 countries and I’ve only been to 14 of those. That’s less than a third. Even within countries, there’s enormous diversity: Ethiopia, where I am right now, is the 2nd most populous African country with 110 million people, and it has over 80 languages. Eritrea, where I’m originally from, is next door. It's about 10% the landmass of Ethiopia and over 80x smaller than the USA but still has 9 official languages – I only speak one of them!
So I’ve been feeling very daunted about this talk because I want to make a fair and dignified representation and…I want to do it in 10minutes. Luckily, I had an epiphany: the inclusivity revolution, that initiatives like 100 Brilliant Women in AI Ethics feed, means I can worry less about the burden of being the token one who has to speak on behalf of all Africans or African women, or all of anything. I can be reassured that I’m not alone and I don’t have to know all the answers and that’s a good place to start.
All I can do is share my story and my reflections on the ways that AI ethics matters in particular to Africa, and why Africans matter to the AI ethics debate. I do that not with the intention of drawing any firm conclusions – quite the contrary – this is an invitation to hear your stories and reflections on Africa in this context.
So, my story:
As I said, I’m originally from Eritrea. I was born in Ethiopia but moved to the UK with my family as a child refugee, escaping the military Derg regime at the time.
After studying and working in the UK, I moved back to Africa – not just to escape the British weather, but because I believe passionately that Africans have both the opportunity and responsibility to shape the continent’s future. I also believe that the best way to do this is to contribute our talents, share our stories, and to collaborate with everyone committed to the cause – African and non-African alike. This cannot be an exclusive venture, and to be included as equals in the tech revolutions, we must be inclusive in our efforts.
I have been living and working across Africa for over a decade – and for most of that time, I’ve been working with African governments through the Tony Blair Institute for Global Change.
The Tony Blair Institute is a non-profit whose mission is to support political leaders and governments to build open, inclusive and prosperous societies – and to do that in an increasingly complex, inter-dependent and technologically-driven world.
I work in our Africa Governance Advisory team – I’m currently based in Addis Ababa, my city of birth.
My primary role is to support the Ethiopian Ministry of Innovation & Technology to implement the National Digital Strategy which was approved by the Council of Ministers in June 2020. I was privileged to lead the team from the Institute that helped draft that strategy in partnership with the Ministry, many government and private stakeholders and also in partnership with the Oxford University’s Pathways for Prosperity Commission – a specialist commission that, like my organisation, was grappling with the big questions around the 4th Industrial Revolutoin and the risks and opportunities it presented for developing countries
My team also recently partnered with the AI4Good foundation to draft Ethiopia’s national AI policy. This will guide the work of the Ethiopian Government’s AI institute which was inaugurated in Sept this year.
If you’re surprised that Ethiopia has a Government AI institute, it’s okay – many of my Ethiopian friends are too!
In addition to my work in Ethiopia, I am part of a global team at the Institute that is responding to the exponentially growing demand from Governments on how to utilise technologies like AI for inclusive development.
My experience professionally and personally informs my reflections on why Africa needs greater representation within the AI ethics debate and why greater inclusion of Africans is critical to that debate and to the continent’s future. I will reflect briefly on three big themes that I think really matter in particular:
1. My first big theme is: The impact of technologies like AI on traditional economic development pathways
That traditional development pathway has typically involved the upscaling from low-grade labour intensive manufacturing to more complex manufacturing and then the shift to services as well as policies that feed innovation to enable the economy to go further up the development latter. This pathway depends heavily on global manufacturing utilising the cheapest labour they can find to kick start the process and get the poorest economies on the first couple of steps up the development ladder.
While this traditional pathway has involved exploitation and huge inequalities, it has also lifted millions out of poverty. The Garment and Business Process Outsourcing sectors, in particular, have economically empowered women in developing countries.
The thing is, new technologies – like robotics, machine learning, big data, and 3D printing – are narrowing that traditional pathway. Many African countries have really just begun competing with South East Asia with respect to cheaper labour, but now the cheapest labourers are sophisticated machinery or customer service bots. It makes sense to have your production as close as possible to the most lucrative markets – or where energy is reliable and cheap – so few African countries can really compete in this new global context.
I want to be very careful not to feed the ‘dark continent’ narrative, but I don’t want to deny or gloss over the fact that Sub-Saharan Africa is disproportionately impacted by extreme poverty.
However, while this threat to economic development as we know it is real, technologies like AI create significant new and unique opportunities for the continent.
The African Union ratified its Continental Digital Strategy in February 2020 and it states:
Tech hubs are growing across the continent – Africans find themselves with unprecedented access to the world via their phone. I don’t think you can walk through the city centre in Nairobi, Lagos, Johannesburg or Kigali without bumping into an African who believes they are a Black Bill Gates in the making – to quote another Brilliant woman.
How can we ensure that these new economic opportunities are fair and inclusive, and that leap frogging won’t increase the economic divide? African women are severely under-represented in these tech hubs and disproportionately represented in the sectors that traditional economic development depends on – what can we do about that? Governments everywhere are struggling to regulate and fairly tax global tech companies, yet Africa is least able to absorb the loss in tax revenue and, while expertise is growing, governments have limited capacity and resources to regulate such firms and the data that is being collected about their citizens.
2. My second theme is: Identity
According to the African Union strategy again: approximately 542 million people in Africa do not have a foundational identification and therefore are “invisible.” Again, women are disproportionately impacted and excluded.
A lack of a legal identity prevents you from claiming your rights as a citizen, your right to your land and to vote, your ability to travel, and prevents institutions recognising your family.
In particular, a means of identity is critical for refugees and internally displaced people.
According to the Internal Displacement Monitor, Africa hosts over one-third of the global forced displacement population.
Immigration and the plight of refugees is naturally close to my heart.
Digital Identity is a key pillar of the AU Continental Digital strategy for good reason. Organisations like ID4Africa and the UN’s Economic Commission for Africa see the potential of digital identity to empower citizens; to deliver better public services; to facilitate economic productivity; and enhance the transparency of institutions. If policies are harmonised across the continent, it can unleash inter-Africa trade and facilitate peace and security at an unprecedented rate.
But all that is based on the premise that Digital ID is done right. What does that mean? Who defines what is right? And what can we do to empower both Governments and citizens to design respectful, safe and effective systems?
3. My final theme is really two in one: Representative Knowledge and its associated Power:
There are many institutional barriers to Africans doing AI research and having African knowledge in this field treated as valid and worthy of global attention. Amongst these barriers are limited or poor quality educational opportunities in specialist fields like AI, a lack of funding for research, and not being able to get a visa to study or attend global conferences. In 2019, over 100 Africans were denied a Canadian visa so they could attend The Neural Information Processing Systems conference. This happened in spite of planned workshops around ‘ML for the developing world’ and ‘Black in AI’ which in fact featured Brilliant women from the Summit’s previous lists.
Nevertheless, I’m optimistic because I see the justifiable growing investment that governments, private sector companies and individual Africans are making .
The University of Lagos launched Nigeria’s first AI hub in June 2018; Google opened its first AI research centre in Accra, Ghana in 2019; IBM’s Thinklab is in Nairobi & Jo’burg – just to mention a few examples.
This is a beginning, but until Africans are curating and owning their data solutions; until they are central in the process of defining the challenges and opportunities that they care about; until we can invest in our own knowledge and be committed to address the inequalities within our own communities, we will not realise the potential of AI and benefit from an ethical approach. We risk remaining on the periphery, as mere recipients of a world view that has historically excluded us.
I repeat my belief in collaboration but, for that to be done as peers, African knowledge, research and voices in AI need nurture and recognition. The experience of Timnit Gebru no doubt has lessons for all those committed to the recognition and nurture of African women in this field.
Again, that’s why I’m so appreciative of this Summit and everything it stands for.
Thank you for your attention – and I look forward to continuing to learn from you.