Good afternoon, ladies and gentlemen. It is my great honour to speak at the World AI Conference session under the theme “AI+ Action: New Roles for Women.” My name is Jingbo Huang, Director of the United Nations University Institute in Macau. I would like to first introduce my organization, as you might not have heard about it.
United Nations University: UNU is a knowledge organization and a UN think tank focused on teaching, learning and generating policy-relevant research related to global pressing issues. UNU has 13 institutes in 12 countries, covering 5 continents. UNU was established in 1975, and next year it will celebrate its 50th anniversary.
UNU Macau: UNU Macau is one of 13 UNU institutes, and the only one specialized in digital technologies and SDGs. Over the past 3 decades, UNU Macau, which has been called the “West Point” of software technology, has trained thousands of researchers all over the world. In the last decade, the Institute has transformed its portfolio toward digital technology and SDGs, in particular AI. We have a few job openings for AI Researchers. Please feel free to apply or help us spread the word.
The theme of gender and technology has been a priority area of our research and training. We firmly believe that women, together with men, can transform our collective future to make it more secure and sustainable. Women are major driving forces to advance SDGs, especially now when the world falls far short of the SDGs targets.
According to the recently launched Sustainable Development Goals Report for 2024, only 17% of the SDGs targets are on track, and more than one-third of the targets have even stalled or regressed. For example, 1 in 10 people globally still faces hunger, and an additional 23 million people have been pushed into extreme poverty due to the lingering impacts of COVID-19 compounded by conflicts, climate shocks and economic turmoil. The world is severely off track to realize the 2030 Agenda.
If we analyze SDG targets specifically related to women, we observe that there are still significant gender digital divides in digital access, skills and leadership.
Globally one in three people are still offline. Worldwide, 70% of men use the internet compared to 65% of women, with 244 million more men than women online in 2023.
In terms of women in leadership, as of January 1st, 2024, women only occupied 26.9% of seats in national parliaments. While women hold 40% of global employment, as of 2022 we only occupied 27.5% of managerial positions--the same share as in 2016. The world hasn’t made any progress in improving the representation of women in leadership positions in 6 years. If we continue at the current rate of change, it will take 176 years for women to occupy 50% of global managerial roles.
How can we improve gender digital divides? How can we unlock the potential of women so the SDGs will be accelerated? UNU Macau has dedicated the past 8 years to these queries. We ask ourselves two sets of questions:
1. In relation to the current situation of women in the digital world: What are the digital gender divides? What contributes to these divides?
2. In relation to the conceptualization of the kinds of interventions we can provide to digitally empower women: What policy interventions at national and international levels can help address these issues?
EQUALS: A few years ago, UNU, together with ITU, UN Women, GSMA and ITC, founded the EQUALS global partnership to bridge the digital gender divide. UNU Macau was the research lead and directed the generation of the inaugural report in 2019 by surveying the landscape of knowledge on digital gender equality. The research report "Taking Stock: Data and Evidence on Gender Equality in Digital Access, Skills, and Leadership" incorporated contributions from 30 multidisciplinary researchers specialized in digital gender inequalities.
The report highlighted some persistent barriers that women face in accessing digital technologies, developing ICT skills, and assuming leadership roles in the tech industry. These barriers include inadequate infrastructure availability, financial constraints, safety and security concerns, and socio-cultural biases. To reduce these barriers, we need multi-stakeholder collaboration among members of academia, the private sector, governments and NGOs.
In addition, our study revealed some counter-intuitive findings. For instance, the report showed that girls from countries in Western Asia perform better than boys in math. As an example, in Saudi Arabia, girls performed 40% better than boys. The result challenged existing stereotypes relating to girls in STEM and showcased areas of progress. The EQUALS report provided an excellent planning data resource for policymakers, practitioners, and researchers interested in solving gender inequality challenges. One important lesson we learned is that we need to work for and with women to bridge the digital gender divide.
AI biases for women: Right before the COVID pandemic, gender biases in AI piqued our interest. A study by UC Berkeley analysed 133 AI systems across different industries and found that 44% of them showed gender biases, taking the forms of discrimination, stereotyping, exclusion and prejudice.
A Turkish artist prompted a Generative AI to write a story about a doctor and a nurse. The Generative AI automatically decided that the doctor was a he, and the nurse was a she.
The gender biases in AI probably came from biased data points. But aren’t these data points reflections of the society we are living in? Discrimination, stereotyping, and exclusion are exhibited in human resources, financial services and government services.
ITU Report: In 2021, UNU Macau began working with ITU to identify and mitigate the risks of gender biases in AI. Our main focus centred on investigating gaps and opportunities related to gender biases of AI in Thailand, Malaysia, Indonesia, and the Philippines, with the objective of recommending gender-sensitive policy development.
Our research first identified significant gender gaps in STEM education and labour force participation. In the Philippines, for instance, women constitute less than 30% of technology graduates and face substantial wage gaps compared to their male counterparts. Gender may be mentioned in many policies, but is not explored in depth as a key risk factor in AI systems. In implementations, there has not yet been an effort made towards generating gender impact assessments.
Based on this research, our recommendations to bridge these gaps are 1. forming government committees dedicated to assessing AI's societal and gender impacts, 2. creating national guidelines for gender-sensitive AI development, 3. setting safe AI standards that include women and address de-biasing at each stage of the AI lifecycle and 4. Investing in women in AI networks and training programs. These initiatives aim to broaden the benefits of AI to women and marginalized groups, fostering a more inclusive technological landscape.
After identifying the four types of gender biases in AI, namely discrimination, stereotyping, exclusion and insecurity, we then focused on women, peace and security and AI. Drawing from our research in Southeast Asia, we concluded that the current AI systems pose security risks to women, particularly in relation to online harms such as cyberbullying and harassment, dis- and misinformation and privacy. As AI will add 1 trillion USD to the GDP of Southeast Asian countries by 2030, the gender risks in AI have to be addressed at regulatory and policy levels. However, not all countries are aware of the gender risks in AI.
UN Women research: Last month, following more than 2 years of research, UNU Macau and UN Women Asia Pacific launched two research reports on “Women, Peace and Cybersecurity” in Southeast Asia.
In this study, we examined AI and women’s peace and security by categorizing AI technologies into 3 areas - AI for peace, neutral AI, and AI for conflict. We then identified gender risks in each category and provided mitigating measures. For example, Neutral AI, such as social media and generative AI tools, are fed with misleading information as often as 80% of the time, which further propagates false narratives. For these cases, we recommended leveraging AI for fact-checking to counter the spread of disinformation.
The main policy recommendations we proposed are: 1. encouraging inclusive and gender-responsive AI design. 2. Raising awareness and strengthening capacity-building among key stakeholders. 3. Calling for a shift from techno-centric to human-centred approaches to promote cyber resilience and digital gender security.
Over the years, we have gathered substantial evidence related to gender and AI through our research, which we used as our basis for developing capacity-building programmes for the UN, governments, and our community in Macau.
UN Women Training: At the UN level, we recently organized a “train the trainers” program in Bangkok. Building upon the research findings of the women, peace and cybersecurity report as well as our AI research, we trained gender equality advocates to strengthen their skills in digital security and cyber resilience.
AI Conference: This April in Macau, we convened the UNU Macau AI Conference, which brought together nearly 500 participants from over 30 countries to engage in 28 parallel sessions. Some sessions included gender angles, and we dedicated an entire session to women and AI. The conference report is out, please feel free to read it.
Training in Macau: Within our community, every year, we are invited by the Women’s Association of Macau to teach in their International Affairs Training Programme. It is so rewarding when the women participating in the Programme tell us that our training has impacted their future career planning.
Coaching: In addition to training, I personally enjoy empowering women’s career development. In my spare time, I am a certified career coach and have been coaching people both inside and outside of the UN system. Among my coachees, 90% are women, and many are working in the AI or tech fields. The female coachees impress me with their resilience, courage, wisdom, hard work and love: love for their families, for their work and for their lives. They are mothers, wives and daughters as well as leaders and managers of their organizations. They want to excel in every role they play.
One female coachee used to work in a top tech company, and held a managerial position. Just as her career was picking up speed, her husband changed jobs and locations. She decided to follow. In moving to a new country, she had no choice but to start from 0. But thanks to her creativity and persistence, she was able to continue to work in the AI field that she’s passionate about, by initiating a new business. Then her husband moved again, and she followed along again. This time, she launched a new career in a new location while still remaining in the AI field. When I commended her on her achievements, she smiled and said, “I feel that I am dancing in chains.”
My wish is that one day, all women will be free from chains - whether digital or societal. I hope that we will work together to continue to bring UN values into digital technology, ensuring that all women will enjoy the opportunities to meaningfully participate, lead and thrive.
I am grateful for the opportunity to share our stories with such an esteemed and dynamic community. I look forward to learning from you and collaborating with you to contribute to a more secure, equal and sustainable future for all.
