Reports
Women’s Rights and AI : A New Threat?
- Women's Rights Team
- 1 July 2024
The rapid advancement in generative Artificial Intelligence (AI) technology is marked by frequent launches of new large-scale models, heralding a transformative era in human work dynamics and communication (Hacker, 2023). However, Generative AI poses numerous risks and challenges. Ethical concerns and potential social exclusion are exemplified by the amplification of societal biases, discrimination, and inequities, notably in fields such as recruitment (Raub, 2018). Another significant risk involves privacy and data protection, encompassing lawful and compliant data collection, processing, and management of sensitive personal information.
Gender inequality persists globally, and AI reflects, and perpetuates these biases ingrained in our society. According to UN Women (2024), despite increasing global internet access among women, only 20 percent of low-income countries are connected, contributing to a significant gender digital divide that is mirrored in AI’s data biases. Research conducted by the Berkeley Haas Center for Equity, Gender, and Leadership examined 133 AI systems across various sectors, revealing that approximately 44 percent exhibited gender bias, with 25 percent displaying both gender and racial biases (Smith & Rustagi, 2021).
While technological progress is generally beneficial, enhancing crucial aspects of human life and boosting societal productivity, it also introduces complexities. AI’s advancement towards a more interconnected world can inadvertently reinforce existing surveillance systems and power dynamics, often in subtle ways (Tacheva & Ramasubramanian, 2023).
Zaki and Meira (2014) state that AI relies heavily on continuous surveillance and data collection to update its systems and understand individuals’ roles and positions. Algorithms are not only dependent on data but can also generate it, making the quality and fairness of algorithmic operations highly contingent on the integrity of the data used. Biased or inadequate data inputs can lead to algorithmic discrimination and compromise their overall reliability (Cheng & Liu, 2024).
Furthermore, in everyday life, there is a common assumption that women bear a greater responsibility for their personal behaviour. They are frequently viewed through roles of, for instance, a woman, wife, mother, or sister. Due to longstanding perceptions of inferiority and vulnerability, women often experience reduced privacy in personal matters, making them more susceptible targets. This perception sometimes justifies accusations against them, leading to their victimisation. These gender norms of behaviour extend into cyberspace, exacerbating the risks they face (Singh Isser, Navneet & Raj, 2024).
Addressing gender bias in AI necessitates prioritising gender equality throughout the development and conceptualisation of AI systems. This entails scrutinising data for biases, ensuring representation of diverse gender and racial experiences in data sets, and fostering diverse and inclusive teams within AI development (UN Women, 2024).
Gender inequality persists globally, and AI reflects, and perpetuates these biases ingrained in our society. According to UN Women (2024), despite increasing global internet access among women, only 20 percent of low-income countries are connected, contributing to a significant gender digital divide that is mirrored in AI’s data biases. Research conducted by the Berkeley Haas Center for Equity, Gender, and Leadership examined 133 AI systems across various sectors, revealing that approximately 44 percent exhibited gender bias, with 25 percent displaying both gender and racial biases (Smith & Rustagi, 2021).
While technological progress is generally beneficial, enhancing crucial aspects of human life and boosting societal productivity, it also introduces complexities. AI’s advancement towards a more interconnected world can inadvertently reinforce existing surveillance systems and power dynamics, often in subtle ways (Tacheva & Ramasubramanian, 2023).
Zaki and Meira (2014) state that AI relies heavily on continuous surveillance and data collection to update its systems and understand individuals’ roles and positions. Algorithms are not only dependent on data but can also generate it, making the quality and fairness of algorithmic operations highly contingent on the integrity of the data used. Biased or inadequate data inputs can lead to algorithmic discrimination and compromise their overall reliability (Cheng & Liu, 2024).
Furthermore, in everyday life, there is a common assumption that women bear a greater responsibility for their personal behaviour. They are frequently viewed through roles of, for instance, a woman, wife, mother, or sister. Due to longstanding perceptions of inferiority and vulnerability, women often experience reduced privacy in personal matters, making them more susceptible targets. This perception sometimes justifies accusations against them, leading to their victimisation. These gender norms of behaviour extend into cyberspace, exacerbating the risks they face (Singh Isser, Navneet & Raj, 2024).
Addressing gender bias in AI necessitates prioritising gender equality throughout the development and conceptualisation of AI systems. This entails scrutinising data for biases, ensuring representation of diverse gender and racial experiences in data sets, and fostering diverse and inclusive teams within AI development (UN Women, 2024).
✨ AI summary
The rapid advancement in generative Artificial Intelligence (AI) technology is marked by frequent launches of new large-scale models, heralding a transformative era in human work dynamics and communication (Hacker, 2023). However, Generative AI poses numerous risks and challenges. Ethical concerns and potential social exclusion are exemplified by the amplification of societal biases, discrimination, and inequities, notably in fields such as recruitment (Raub, 2018). Another significant risk involves privacy and data protection, encompassing lawful and compliant data collection, processing, and management of sensitive personal information.
Gender inequality persists globally, and AI reflects, and perpetuates these biases ingrained in our society. According to UN Women (2024), despite increasing global internet access among women, only 20 percent of low-income countries are connected, contributing to a significant gender digital divide that is mirrored in AI’s data biases. Research conducted by the Berkeley Haas Center for Equity, Gender, and Leadership examined 133 AI systems across various sectors, revealing that approximately 44 percent exhibited gender bias, with 25 percent displaying both gender and racial biases (Smith & Rustagi, 2021).
While technological progress is generally beneficial, enhancing crucial aspects of human life and boosting societal productivity, it also introduces complexities. AI’s advancement towards a more interconnected world can inadvertently reinforce existing surveillance systems and power dynamics, often in subtle ways (Tacheva & Ramasubramanian, 2023).
Zaki and Meira (2014) state that AI relies heavily on continuous surveillance and data collection to update its systems and understand individuals’ roles and positions. Algorithms are not only dependent on data but can also generate it, making the quality and fairness of algorithmic operations highly contingent on the integrity of the data used. Biased or inadequate data inputs can lead to algorithmic discrimination and compromise their overall reliability (Cheng & Liu, 2024).
Furthermore, in everyday life, there is a common assumption that women bear a greater responsibility for their personal behaviour. They are frequently viewed through roles of, for instance, a woman, wife, mother, or sister. Due to longstanding perceptions of inferiority and vulnerability, women often experience reduced privacy in personal matters, making them more susceptible targets. This perception sometimes justifies accusations against them, leading to their victimisation. These gender norms of behaviour extend into cyberspace, exacerbating the risks they face (Singh Isser, Navneet & Raj, 2024).
Addressing gender bias in AI necessitates prioritising gender equality throughout the development and conceptualisation of AI systems. This entails scrutinising data for biases, ensuring representation of diverse gender and racial experiences in data sets, and fostering diverse and inclusive teams within AI development (UN Women, 2024).
Gender inequality persists globally, and AI reflects, and perpetuates these biases ingrained in our society. According to UN Women (2024), despite increasing global internet access among women, only 20 percent of low-income countries are connected, contributing to a significant gender digital divide that is mirrored in AI’s data biases. Research conducted by the Berkeley Haas Center for Equity, Gender, and Leadership examined 133 AI systems across various sectors, revealing that approximately 44 percent exhibited gender bias, with 25 percent displaying both gender and racial biases (Smith & Rustagi, 2021).
While technological progress is generally beneficial, enhancing crucial aspects of human life and boosting societal productivity, it also introduces complexities. AI’s advancement towards a more interconnected world can inadvertently reinforce existing surveillance systems and power dynamics, often in subtle ways (Tacheva & Ramasubramanian, 2023).
Zaki and Meira (2014) state that AI relies heavily on continuous surveillance and data collection to update its systems and understand individuals’ roles and positions. Algorithms are not only dependent on data but can also generate it, making the quality and fairness of algorithmic operations highly contingent on the integrity of the data used. Biased or inadequate data inputs can lead to algorithmic discrimination and compromise their overall reliability (Cheng & Liu, 2024).
Furthermore, in everyday life, there is a common assumption that women bear a greater responsibility for their personal behaviour. They are frequently viewed through roles of, for instance, a woman, wife, mother, or sister. Due to longstanding perceptions of inferiority and vulnerability, women often experience reduced privacy in personal matters, making them more susceptible targets. This perception sometimes justifies accusations against them, leading to their victimisation. These gender norms of behaviour extend into cyberspace, exacerbating the risks they face (Singh Isser, Navneet & Raj, 2024).
Addressing gender bias in AI necessitates prioritising gender equality throughout the development and conceptualisation of AI systems. This entails scrutinising data for biases, ensuring representation of diverse gender and racial experiences in data sets, and fostering diverse and inclusive teams within AI development (UN Women, 2024).
Share this report
Facebook
X
LinkedIn
WhatsApp
