How queer is Artificial Intelligence?
AI is playing an increasingly important role in our lives, but that is not without danger. Eduard Fosch-Villaronga at eLaw warns that AI does not pay enough attention to the LGBTQ community.
Eduard Fosch-Villaronga, Associate Professor at eLaw specialising in AI, robotics, and the law, emphasises diversity and inclusion's critical role in AI development. His insights and research on this topic are featured in Winq thanks to Erwin Vogelaar, shedding light on the potential risks associated with biased AI systems, especially about the LGBTQ community.
AI and its binary thinking
Fosch-Villaronga highlights the limitation of binary thinking in AI development, where systems are often designed to work in a black-and-white manner - either they function correctly or don't. According to Fosch-Villaronga, this oversimplified approach neglects the rich spectrum of human diversity. He draws attention to the issue of gender classification in AI systems, noting that the commonly used binary parameter of male or female overlooks the existence of the intersex community. Moreover, he points out the distinction between gender and sex, emphasising that gender is subjective and may not always align with external characteristics, leading to potential errors in machine predictions.
The consequences of biased AI systems become particularly concerning when AI is involved in life-defining moments, such as occurred in the childcare benefits affair in the Netherlands. Fosch-Villaronga underscores the increasing integration of AI in healthcare for tasks like diagnoses and treatment planning. In cases where an individual's gender identity changes, traditional AI systems may struggle to adapt, potentially resulting in misgendering. In a medical context, misgendering could lead to misdiagnoses, seriously threatening individuals' safety.
Fosch-Villaronga's research and insights serve as a call to action for the AI community to address the biases inherent in the data used to train these systems. Achieving diversity and inclusion in AI is not just an ethical imperative; it is essential for creating AI applications that are accurate, fair, and safe for all individuals, regardless of their gender identity or other characteristics.
Diversity, Robots, and AI
Eduard Fosch-Villaronga is committed to advancing the knowledge in the field of Diversity, Robots, & AI. Within that topic, Eduard co-chairs the Gendering Algorithms initiative at Leiden University, together with Tessa Verhoef from the LIACS department, a project aiming to explore the functioning, effects, and governance policies of AI-based gender classification systems.
Erwin Vogelaar and Winq
Erwin Vogelaar is a Dutch freelance journalist, who writes about technology for newspapers, magazines and websites. Winq is the biggest queer lifestyle magazine in the Netherlands.