Good AI requires women in the data, in the workforce, and at the table
When urban planners design a city of the future, they make assumptions about how future citizens will behave. Those assumptions tend to reflect available data, as well as the ideas, perceptions and biases of the planning team.
“We’re using digital twins to understand behaviour patterns for predictive analysis in smart cities,” explains Pico Velasquez, founder and CEO of VIIRA, a design-technology company focused on the built environment.
But without broad and accurate data, the best-intentioned urban plans may still overlook parts of the population, such as women and minorities.
In effect, women can find themselves sidelined in decisions about future buildings, public transport, and other city infrastructure if artificial intelligence (AI) models favour typical male preferences or overlook female needs.
“The mobility and transport preferences of women differ from those of men,” Velasquez noted during a panel discussion on standards for inclusive AI organized by the International Telecommunication Union (ITU) on 17 October.
However, community participation and transparent data practices can help overcome gender-biased planning for cities of the future, she said.
The panel formed part of the Network of Women (NoW) in ITU standardization special event during the recent World Telecommunication Standardization Assembly (WTSA-24) in New Delhi, India.
Diverse data sets will also improve broader AI development, agreed other panellists.
Sandra Maximiano from Portugal’s telecom regulator, ANACOM, emphasized the need to tackle pervasive prejudices, statistical bias, and stereotyping. “90 per cent of men and women exhibit bias against women,” she said.
Calls for inclusive AI
Alessandra Sala, Senior Director of AI & Data Science at Shutterstock, called for transparency, accountability, and international standards to ensure diverse representation in AI data. “AI models have to be audited for strategic measures of bias,” she said.
Ways are also needed to define when problematic biases in AI become outright unacceptable or dangerous.
Ebtesam Almazrouei, founder and CEO of AIE3, added: “We need to embed red lines in AI development to ensure inclusive AI systems.”
Participation in the standardization process is crucial, especially in the age of AI.
“In this new tech era, ensuring greater, and meaningful, participation will help in the development of inclusive, sustainable standards,” said Rim Bellhassine-Cherif, Chair of the Network of Women in ITU-T and Chief Innovation Strategy Officer at Tunisia Telecom, said
UN Women India country representative and panel moderator Susan Ferguson concluded the session with a reminder: “AI is for humans. It is about us using it and us benefitting from it.”
Imbalance in the industry
Balanced workforces can help create digital products that meet everyone’s needs. Yet the tech industry remains male dominated. Women remain underrepresented even in new fields like AI.
According to the World Economic Forum, only 22% of AI professionals globally are female. The gender gap raises concerns about the diversity of perspectives in AI development and whether the resulting technologies are equitable.
Across the global ITU community, female representation remains below targets, highlighting the necessity for assertiveness and mutual support to overcome these barriers.
Women’s participation for this WTSA stood at 26 per cent, below the “NOW4WTSA-24” campaign target of 35 per cent. Still, the proportion of women’s leadership in national delegations and WTSA committees is growing.
Women’s representation can vary widely between countries and regions. Madhu Arora, a member of the Technology/Digital Communications Commission at the Department of Telecommunications in India, noted that Europe led other world regions with 34 per cent women at WTSA-24.
Seizo Onoe, Director of the Telecommunication Standardization Bureau at ITU, reaffirmed the organization’s commitment to gender equality. “We are committed to moving forward together. That is exactly what our Network of Women is about,” he said.
Encouraging women in STEM
“We have to encourage women,” said Burkina Faso’sMinister of Digital Transition, Posts & Electronic Communications, Aminata Zerbo-Sabané, stressing the importance of exposing young girls to technology early on to build their confidence and interest in science, technology, engineering, and mathematics (STEM). “Give them a chance to dream, get careers in STEM in fields that today are just men.”
Women often struggle with accepting leadership roles due to family responsibilities, although they can excel once they take on such roles. Flexible working hours can help manage work and home responsibilities, Zerbo-Sabané added.
“A baby is a real thing. You can’t ignore it,” concurred Neha Satak, Co-Founder and CEO of millimetre-wave wireless solution developer Astrome. But the bigger challenge is to get young girls engaged with tech early.
Satak shared her early fascination with unmanned aerial vehicles (UAVs) and the scepticism she experienced from people around her until she showed them how to get one off the ground. “If you persist, it will happen,” she said.
Researcher Vishnu Ram, a contributor to ITU standardization work on AI and machine learning, said he had learned to address his own innate biases as a tech expert and mentor. Young women must publish their tech research and raise their voices in the industry, he said.
Places at the table
ITU Secretary-General Doreen Bogdan-Martin agreed: “If there is no seat at the table, you have to bring your own chair.”
The Beijing Declaration, a resolution adopted by the United Nations in 1995, outlines principles to ensure equality for men and women both in principle and in practice.
ITU – the UN agency for digital technologies – has adopted gender equality and mainstreaming resolutions in all sectors of its activities.
Bogdan-Martin stressed the need for mutual support among women in the field.
“We have a lot more work to do, and I’m confident we will do it,” she said.
Credit photos: ©ITU/D. Woldu
