Today in the midst of multiple global social and economic crises and a pandemic, we honour Ada Lovelace – tech visionary and Victorian badass – in order to celebrate and spur more numbers and different numbers of women designing the technologies of our future.
It is imperative that we systematically expand opportunities for women to solve large-scale social problems by enlarging who is included in the problem definition and solution design of cutting-edge technologies, especially Artificial Intelligence.
Newly strategic thinking is needed inside traditional tech and government structures. This is our daily policy and advocacy struggle at the <A+> Alliance for Inclusive Algorithms.
It is widely acknowledged that AI has a:
- Lack of diversity in the discipline, with a majority of men creating, designing, implementing and deploying the systems. Instead of increasing, the numbers of women have been declining, over the last years.
- Lack of diversity in the data and in the algorithms themselves that provide recommendations of news/ movies /videos /etc, and aids in decision- making of credit lines /bail /job recruitment /government allocation /medical treatment /etc leading to a credibility revolution,
- Lack of diversity in the type of organizations creating the cutting-edge AI systems of today: scientific production of AI is led by a handful of US or Chinese companies, not by universities, research centres, non-profits or grassroots organizations.
All this is adding up to the gender roles very slowly being removed from the physical world being newly hardwired with old stereotypical conceptions and associations of gender, race, and class into new models of AI and Automated Decision Making.
According to the World Economic Forum, only about 22 per cent of all AI developers are women; most of them based in the top 10 richest countries. (These figures do not count the unsung, overlooked and underpaid female data cleaners, content moderators and gig workers from the global south). They all struggle to reach (and to be) decision-makers and investors. Equally troubling and even more unacceptable is the lack of policy and standards to accelerate equality in the algorithms themselves, and policy to impede the unequal futures lurking within old data and old models which, if not dismantled, will become hardened and evolve into a more effective Patriarchy 2.0.
Many of the investments and infrastructures of the future will have tech solutions at their core, so embedding affirmative actions at all stages of the processes is imperative.
In the next months, the world’s leaders will take important steps to address the converging crises: from allocating millions in economic relief funds to investing in new health systems. Entire regions are already on a path to a green transition. Many of the investments and infrastructures of the future will have tech solutions at their core, so embedding affirmative actions at all stages of the processes is imperative.
At the <A+> Alliance for Inclusive Algorithms, we propose a holistic approach to solve the tech exclusion problem from its roots:
We call for taking dynamic steps to include an intersectional variety and equal numbers of women and girls in the creation, design, and coding of the technology we want. Beyond encouraging more investment and better strategies for women and girls to gain skills: we want diverse women proactively involved in all AI processes, shaping the technology that affects every part of our lives. We propose the vigorous participation of women and girls, from settlements to cities, from rural environments to primary and secondary schools and universities, better informing the design and implementation decisions of government officers and companies, engineers and management teams miles away from societies they are trying to serve and the problems that they are trying to address. At the <A+> Alliance for Inclusive Algorithms, we believe that there is not only room for, but urgent need for, women and girls’ expert frontline experience and input, participation and co-creation, at each and every stage of AI development and deployment.
We call for Affirmative Action for Algorithms that correct for bias from the moment of their conception. And algorithms that solve the problems needed on the ground with problem definition and solution design sourced from the frontline. We advocate for the adoption of guidelines that establish fairness, accountability and transparency for algorithmic decision making (ADM) in both the public and private sectors, opening up and debiasing the algorithm Black Box.
You do not need a math degree to have lived and understood the systemic fixes that need to be escalated in society, to include more, and to serve better. By combining the knowledge on the ground from the feminist frontline with more diverse and inclusive teams composed of computer scientists, machine learning experts, data scientists and mathematicians, an alliance can be created for tackling and actively correcting for systemic gender and race inequalities. This, in our view, is a better avenue than taking old systems and old assumptions, optimizing them for efficiency and digitizing them with old biases at scale; and only then -maybe- correcting for bias after harm has been caused.
Given the pace at which AI and other Automated Decision-Making systems are being deployed we need to seed the future we want with inclusion, multi-disciplinarity, and gender, race and class diversity. We are ready to pilot that future, seizing the urgent now of COVID-19 response work firmly in the context of the equally urgent need to create more democratic, new systems so that we all can thrive.
Our proposal is reasonable and possible to pilot: open the room and seat women and girls at your COVID-19 table where you are deciding the next tech interventions, including AI interventions, to solve the current crises. Test a participatory model. See the difference it makes in the outcomes that create the collective future we deserve. See and seize women and girls’ potential. Repeat the process again and again. Once a space for innovative thinking is opened, the world will see one million Adas emerge. But only if an ecosystem for their ideas to flourish is created now.
This piece is signed by Renata Avila as a member of <A+> Alliance's advisory board, along with Caitlin Kraft-Buchman, Nuria Oliver, Elisa Celis and Nanjira Sambuli.
*<A+> is a multidisciplinary, diverse, global and feminist coalition of expert practitioners, academics and activists working to create and apply Affirmative Action Algorithms (<A+>) that not only detect, but correct, gender bias in Artificial Intelligence and Automated Decision Making. Formally launched at the 2019 Internet Governance Forum, <A+> was named one of Fast Company’s 2020 World Changing Ideas in AI & Data in June 2020, and is a leader of the UN Women Generation Equality Action Coalition for Technology and Innovation.