A Conversation With Women in AI
By Claudia Pompa* and Kim Ochs*
Earlier this month, three contributors to the Leading with AI newsletter took part on a panel about the role of women in Artificial Intelligence (AI) as part of Women in Tech Africa 2021. The conversation was a collaborative effort between Women in Tech Africa and GlobalX Innovation Labs as part of GlobalX's Inclusive AI initiative to encourage more women to get into AI. The event was organised and hosted by Sylvia Mukasa, CEO of GlobalX Investments Ltd/GlobalX Innovation Labs. Panellists included: Nomso Kana, based in South Africa, entrepreneur in the tech space as well as Presidential Fourth Industrial Revolution (4IR) Commissioner; Katharina Miller, based in Madrid, founding partner of 3C Compliance; and Claudia Pompa, based in London, Founder and Managing Director of Consulting for Growth Group (C4G).
Excerpts and highlights from the panel discussion are summarised below:
AI is expected to contribute US$15.7 trillion to global GDP by 2030 according to research by PwC. It is therefore crucial for Africa to take a share of this piece of the pie. Despite progress in the sector, women continue to be under-represented in this space, not just in Africa, but globally. Element AI research shows that globally, only 12% of leading machine learning researchers and; only 18% of authors at leading AI conferences are women.
As Nomso pointed out, ethical AI systems must be inclusive, explainable and have a positive purpose in the responsible use of data, specifically to Africa. With a lack of annotated data, there is currently a lot of bias used in coding because things are built from a Western perspective.
Africa has an advantage as a youthful continent: 60% of the population is under the age of 25. Transformation is underway as African governments are looking at how to use youthful innovation and how they are using technology to solve community challenges. They are using AI and building machine models to understand their own ecosystems. However, there needs to be a path for innovation to come into the public sector, which includes convergence between the public and private sector, as well as partnership - including stakeholders such as youth and women - to formulate policies to govern AI and advanced technologies.
Claudia’s research has focused on the future of work, understanding how work is going to change in the next decade. The research they have conducted indicates that automation technologies, particularly AI, is going to have the highest impact in terms of jobs, wages and skills. In this scenario understanding not only what AI is, but also what are some of its societal impacts, is crucial. Despite how exciting the AI world is at the moment, the reality is the sector is still defined by the “lack of…”: there is a lack of sufficient talent, lack of diversity among the existing talent, lack of sufficient training, lack of digital infrastructure, lack of understanding, lack of regulation, etc. She proposed three top areas to address to bring change to shape the sector and share the societal benefits that AI can provide: focus on the talent pipeline, regulation and funding.
In terms of the talent pipeline, worrying about talent at university level is simply not enough. Talent needs to be nurtured from early ages and we need to provide ample opportunities for young people along their educational pathways to come into contact with different types of technology and to explore those interests. We also need to stop the bifurcation of skills. This refers to the idea that you should pursue either a path on sciences and technology or the social sciences. This is a dangerous way forward because it could easily lead to social tensions, and raise skills and wage disparities. What we really need is a workforce equipped on both fronts, with the technical skills required for an AI future, but also the skills that makes us human – creativity, analytical thinking, empathy, compassion - and able to leverage those interactions.
On the regulation side, after the Frances Haugen revelations about Facebook, we all understand the need to regulate tech companies in general, not only large ones. We need to really think at a societal level about how we use AI, how and who controls AI, and who has ultimate responsibility when things go wrong. We know the market is not going to regulate itself and rather than fight regulation we need to think about smart regulation – regulation that promotes innovation, but at the same time protects the well-being of society in general. However, a huge challenge remains having policy makers who actually understand the technology and how it behaves and what to do about it. To overcome this challenge we need to work together.
Last but not least, we need to address funding by increasing levels and allocating it differently. Funding already available in the AI sector is not necessarily reaching the right people, the right teams or the right geographies. We need to fund more research, more diverse teams, and more women, and we need to fund all of these across geographic regions, not only in Silicon Valley or London. There are amazing ideas and tremendous talent that are not necessarily in these locations.
Katharina noted that Europe has made grade strides with regards to ethics and General Data Protection and Regulation (GDPR), a European law that has to be applied by all individuals and companies living in the European Union as well to those outside of Europe who have relations with those in Europe. The implementation of GDPR has also meant educating Europeans about the use of their data, which has often been described as “the new gold”. However, it is not the data, but rather what companies do with our data that is the gold. Companies create new models, most of the time without our consent, and this is what GDPR wants to protect. In 2016, the EU created a high-level expert group on AI and, in 2019, published its Ethics Guidelines for Trustworthy AI. This document includes ideas on how to make AI more human-centric. The Council of Europe’s Committee on Artificial Intelligence is also working on providing guidelines and publications, with recommendations to address AI and related technologies, such as facial recognition and algorithmic systems, and to address human dimensions of AI technologies.
A challenge remains that law around AI is more dominant in the northern hemisphere and does not effectively include the global south. The power structures in place in the sector – particularly in terms of human resources, financing and regulation - make it easier for the north to create the algorithms, which have been reflected in algorithmic biases. For example, early versions of facial recognition software could not recognise people of colour. To counter this problem, it is important to educate people and create bigger teams to collect data and to create more inclusive algorithms. Regulation will also be important to address this dimension. Dialogues and exchanges across Europe and Africa about issues around privacy, data protection and ethics are needed.
All of the panellists acknowledged the need for more women in the sector, including the need for more role models, and proposed a number of ideas. These included the creation of policies to enable women to tap into this industry, acknowledging that the AI industry not only refers to technical sectors, but can also be used in broader sectors. We need women at all levels, in all sectors and areas. We not only need coders and programmers, but also female doctors who understand AI and how it can be used to treat patients, teachers who understand AI and can then teach the next generation about it, policy makers who can help regulate it, and executives who know how they can use it in their businesses.
In addition, there must be incentives to support diversity and inclusivity in the development of algorithms. We also need to think about the support systems needed for women to participate in this sector. In summary, women in all sectors need to have a seat at the AI table. The topic is too important to have only a few people in remote locations decide the future of so many of us.
*Claudia Pompa is the founder and managing director of the Consulting for Growth. She specialises in issues related to the future of work and workforce development programmes and has extensive experience in digital economies, innovation, economic growth, start-ups and SMEs.
*Kim Ochs has been a researcher and practitioner in the field of educational technology for more than a decade, working with universities, international organisations, private companies, and edtech investors.