As of: February 27, 2024, 8:01 a.m
By: Anika Zuschke
Comments
Press
Split
Women play an important role in AI research.
© Nicolas Bruckmann/DALL-E (AI generated)
As in many technical areas, women are underrepresented in the professional segment of artificial intelligence.
This can have fatal consequences.
Frankfurt – Artificial intelligence is revolutionizing the job market – according to the “Global Gender Gap Report 2023” from the
World Economic Forum,
the concentration of AI talent worldwide increased sixfold between 2016 and 2022.
However, this particularly applies to the male workforce.
Only around 30 percent of AI talent were women in 2022 and the proportion has only increased by four percent since 2016.
According to the UN cultural organization
Unesco,
only around 12 percent of AI researchers worldwide were female in 2023.
Furthermore, women who already work in the AI industry are often ignored or overlooked.
This is shown by an article in the
New York Times
from December 2023, in which 12 people were named AI pioneers - not a single one of them women.
A 2023 analysis by AKAS also found that mentions of Sam Altman (CEO of OpenAI) in AI-related articles were twice as high as the total number of 42 women from the current top 100 list AI Influencer in
Time Magazine
.
Who is responsible for AI – diversity in training data plays a big role
The perspectives and personal experiences of women play a crucial role in the research and design of artificial intelligence.
The reason for this is simple: in order to avoid biases within AI systems, a wide range of lived experiences and needs must flow into the training, design and implementation of AI tools.
In addition to including different ethnicities, socioeconomic backgrounds, professional or age-related experiences, this
of course also includes women's gender-specific perspectives, according to the consulting firm
Deloitte .
This is confirmed by Mia Dand, founder of the global non-profit organization “Women in AI Ethics”, when asked by
IPPEN.MEDIA
: “Women play a crucial role in AI research as they represent half of humanity and all research is based exclusively on based on the priorities and goals of a small, privileged group of one gender, is inherently flawed and inadequate, and will continue to lead to unethical results.”
If we want AI to benefit all of humanity, then it is only logical that AI research should incorporate perspectives from all of humanity.
Mia Dand, founder of Women in AI Ethics
Women in AI research: A lack of integration of women has fatal consequences
“A lack of awareness or concern in the male-dominated tech industry about issues affecting women has resulted in research on current threats such as [...] AI-generated deepfake pornography being significantly underfunded, while at the same time research on future ones “The imaginary existential risks of AI are overfinanced,” Dand continues.
In fact, there are already AI programs that allow users to convert a photo of a clothed woman into a version in which she is naked.
This program only works with images of women, Danielle Keats Citron discovered in her article “The Continued (In)visibility of Cyber Gender Abuse,” which
appeared in the
Yale Law Journal in November 2023.
If you submit a photo of a man to the AI program, it will be converted so that the man receives breasts and female genitals.
The program was trained using actual nude photos of women and has already been used by over 100,000 people.
My news
Traditional company has to close and lays off all employees
When it comes to work-life balance, top economist demands: “We should all learn a lesson from Generation Z”.
German family business files for bankruptcy - 150 employees affected read
Thousands of employees can look forward to a special payment for February's payroll
Next real estate insolvency: Construction crisis is dragging more and more companies into the abyss
Retirement age for those born in 1960: You're still in luck
Discrimination, prejudice, stereotypes – when women are not part of AI research
If the needs and experiences of women are neglected in the design of AI tools, this becomes apparent in the use and sometimes discriminatory results of these tools.
The underrepresentation of women in AI research can also reinforce harmful prejudices and gender stereotypes - or even create new ones.
This problem is evident, among other things, in the healthcare sector: If diagnoses generated by artificial intelligence are only fed with data from men, this can have dangerous consequences for women.
Expert reports: Ways to combat the underrepresentation of women in AI research
But how can this problem be counteracted?
Computer scientist Sheila Beladinejad, who works as founder and president of the organization “Women in AI & Robotics Germany”, has some recommendations: “Research institutions could promote the participation of women in AI research by developing various programs such as public relations and scholarships offer, develop integrative work guidelines and promote women in leadership positions,” she explains to
IPPEN.MEDIA
.
It's important to increase the number of women in the industry, but we also need to ensure they are paid fairly and equally [to men] and given leadership opportunities.
Sheila Beladinejad, founder and president of the organization “Women in AI & Robotics Germany”
She said the industry was now aware of the importance of diversity in its workforce and some attempts had been made to address or correct the inequalities.
“However, progress is slow [...],” explains Beladinejad.
Therefore, “perhaps now is the time to consider stricter consequences for failure to take appropriate action on gender equality in the workplace.”
Whatever the solutions look like, it is of fundamental importance that women are also integrated into AI research and are taken seriously and seen as part of the workforce.
Otherwise, AI tools will continue to emerge that only address the needs of half the population.