by Lionesses of Africa Operations Dept
Without doubt the above is a beautiful picture from the UN Economic Commission for Africa’s recent publication entitled: “Increasing women’s access to capital critical in Africa’s fight against poverty.” Sub-titled: “Women entrepreneurs continue to face greater barriers to accessing financing for their businesses”, see here.
Is it a photo of two friends out shopping for food, or as the title suggests (and our subconscious assumes) a smallholder farmer selling her hard worked (and let us be in no doubt farming is back breaking work) and beautiful fresh produce at the local market? As we move from the Plastic Age (previous ages being Stone, Bronze, Iron, Industrial, Plastic…) and into the Ai age, so such pictures take on a new and important significance.
Here is a (very) basic primer on what it takes to build Ai.
Let’s get one basic fact straight, Ai is not a gift, beautifully wrapped and handed over in a finished state from a higher being to the world wrapped with a pink bow. It was invented, designed and most importantly trained within a male dominated industry. This training is important, because without it Ai cannot learn and work out what to do next. Obviously such a brain is only as good and as accurate as the training and information given. This training takes many forms, but the ‘trainer’ will often fire all available data and millions of images at the ‘brain’ so that its memory and ability to ‘assume’ the next picture or route will start to take hold.
What does this mean in reality for Africa?
Firstly, there is a huge lack of Data on Africa:
Which means if you are building an Ai for Africa, there is simply not enough Data to build the ‘brain’. You have to rely on images.
So to find what images are available we turned to one of the world’s greatest data magazines, The Economist. It is often said that The Economist ‘eats data for Breakfast, Lunch and Dinner’ and with an honorary Lioness in charge such as Editor-in-Chief Zanny Minton Beddoes, we are obviously huge fans.
In their article (here) titled:
“Bias in, bias out
Demographic skews in training data create algorithmic errors.
Women and people of colour are underrepresented and depicted with stereotypes.”
They show charts of the types of images on Google:
(note they used a log scale as the chart would never fit given the huge amount of Marriage/Veil/Lipstick/Brassiere/High Heels images. Photos with marriage of women being 100 times greater than men’s)
So that we are clear - these are the kind of images used to train Ai and if you search for the business/wealth labels, they contain for males: Businessperson/Public Speaking/Management
vs for females: Hmmm….
Note, as they say the share of images labelled ‘Marriage’ was 100 times greater than for men. It is not so much what we are teaching our daughters, but what they are seeing and absorbing that counts. This is also true for Ai…
Followed by:
(please note the log scale again)
Images with light skin: Coach/Rich Person/Cardiologist/Heir/Captain/Director and Coach
vs for those with dark skin: Farmer/President/First Lady (Thank heavens for the Obama’s!) and Travelling Salesman.
Sadly we can all guess the future if all decisions (finance, employment, even personal credit as seen by the gender bias in the Apple Pay Card here) are guided or made by Ai.
This is why we have to ensure that images tell the TRUE story. As Lionesses we can’t simply hide away as Ai will just continue these myths about Women Entrepreneurs - as one of our ‘100 Lionesses’ found out when she applied for a Loan from her Bank and was asked why she did not just ask her husband for the money! This Lioness owns and runs one of the most successful companies in Africa - yet ‘marriage’ was what was foremost in this Bank Manager’s mind - given the images on Google, Ai too will be assuming there is a husband there to help!
Be bold, publish photos of yourself in front of your business, tell the world proudly you are a Lioness, a CEO, a Leader, you as a woman entrepreneur have built this incredible business, have employed people and have brought hope and prosperity to their communities. Don’t allow images such as our title image be the only ones seen on the internet - look again at the Economist charts - ‘Farmer’, ‘Friendship’, ‘Smile’, ‘Happy’, ‘Dress’ these are the labels that Ai sees. Of course this is wrong, but we are in the Data and soon to be Ai age. We have to recognize this. Those who ignored Iron and stayed with Bronze as the world changed, did not last long.
But images alone are not enough.
The HoF tells of the time he introduced himself to a very serious looking person at an international summit, only to be greeted by the comment: “You don’t want to talk to me, I only lend in $10million clips.” Given Melanie had recently provided a very large international bank at the request of MasterCard, 5 Lionesses whose businesses could handle that kind of ticket (we could have given more from our inspirational ‘100 Lionesses’, but were only given 3 days to call, check if interested, gather financials, prepare and send - as you can imagine, not an afternoon’s work!), the HoF was able to point out this fact and only then there was interest - Data is so important!
This person sadly believed that the Lionesses could not be of interest because of ingrained beliefs that women entrepreneurs are only of a particular size and especially in Africa. He did not have the Data and certainly had not done his Research, as the HoF ‘politely’ pointed out!
We have to change the narrative, change the bias before it is too late, find and publish the Data. If not - think it is difficult for Lionesses to get finance before Ai? Sadly, “You ain’t seen nothing yet.”
As the FT said (here): “Africans need to start demanding more… More data won’t fix that by itself. But if governments know…exactly what is going on, they have less excuse to ignore it.” We would add that if Investors knew more about what was going on, they too would not only have less excuses, but would become very interested.
Data4SDGs agree (here): “Data…forms the backbone of decision-making and accountability, but too many people are left out of the data, and too much of the data is outdated or incomplete.”
This is why we have launched (Drumroll please…) Lioness Data. Our Research Director, Linda Zuze, a Phd in Economics, a Zambian and a true Lioness, is determined to drive meaningful data to not only change the world but to lead the world in their understanding of the incredible work being done by Lionesses.
Our first research study, conducted in partnership with New York University and supported by ABSA was published this week and can be found here. Titled: South African Women Entrepreneurs Job Creators Survey - it is a unique survey designed to provide better insights into the contribution being made by women entrepreneurs to job creation in the country. It has some fascinating insights, such as the large percentage of Lionesses who took a pay cut or no pay to keep their employees on during Covid; what Lionesses look for when employing new staff (at only 7% it’s certainly not academic or vocational qualifications!); how they were impacted by Covid; and of course a number of very interesting insights on access to finance.
Please download it, share it, discuss it, talk about it, celebrate it - this is just the start. It is the first in a long line of research publications we shall be bringing out soon, not only to highlight the incredible work our Lionesses do, but most importantly to drive the data and the discussion to deliver a solid foundation for Finance (in all its forms) to truly understand the scale of opportunities available.
This can only come through Data, though Images and through You. For those Lionesses who have helped and are currently helping across Africa and even those who are about to help (but we have not been contacted as yet) with our surveys, many, many thanks --
You have no idea of the absolutely essential service you are doing for all Women Entrepreneurs across Africa now and for The Future.
TOGETHER WE CAN ADD MEANINGFUL COLOUR TO THE MAP.
Stay safe.