AI moderation often fails marginalized groups, especially disabled women and girls, because existing datasets ignore their lived experiences, leading to missed abuse, over-censorship, and reinforced bias. Researchers advocate for co-created, intersectional datasets and contextual AI that understands nuance, intent, and transparent explainable systems to build trust. Addressing data scarcity, bias, and accessibility requires community-driven design, continuous feedback, and survivor-centered policies that prioritize safety, accountability, and wellbeing. In this episode of "Disability Dialogues," Dr. Karen Soldatic and Researcher Glaucia Melo discuss how contextual AI can help detect and prevent technology-facilitated violence against disabled women and girls. They explore a research project that develops AI models using data based on lived experiences to reduce bias and better recognize nuanced online abuse. The conversation highlights the importance of community-driven, inclusive AI, challenges in data collection, and the need for stronger policies and digital literacy to ensure safer, more equitable online spaces for marginalized groups. This episode offers a compelling vision for the future of online safety—one where contextual AI, co-created with marginalized communities, can detect and prevent technology-facilitated violence in all its forms. By centering lived experience, embracing nuance, and demanding accountability, we can build digital environments that are not only safer but also more just and inclusive. Definitions: AI: The development of machine-based systems that can perform tasks typically requiring human intelligence, such as learning, reasoning, problem-solving, and decision-making, often with a degree of autonomy. Ethical AI: Artificial intelligence that adheres to well-defined ethical guidelines regarding fundamental values, including such things as individual rights, privacy, non-discrimination, and non-manipulation. Data sets: A collection of related sets of information that is composed of separate elements but can be manipulated as a unit by a computer. Technology-facilitated violence: behaviours where technology is used to cause virtual and/or in-person harm. Metadata: Dr. Karen Soldatic sits down with researcher Glaucia Melo to discuss technology-facilitated violence against disabled women and girls, and the efforts underway to address and combat this issue. Here are three reasons why you should listen to this episode: Learn about the role of contextual AI in understanding the nuances of online communication. Understand the importance of datasets based on lived experiences to avoid biases in AI. Discover the use of emerging technologies, particularly contextual artificial intelligence (AI), to detect and prevent online abuse. Resources Explore the CERC in Health Equity and Community Wellbeing’s Website Glaucia Melo Website Dr. Karen Soldatic: LinkedIn | CERC Profile Toronto Metropolitan University: Website SSHRC – Social Sciences and Humanities Research Council: Website CIHR – Canadian Institutes of Health Research: Website NSERC – Natural Sciences and Engineering Research Council of Canada: Website Enjoyed this Episode?If you did, be sure to subscribe and share it with your friends! Disability Dialogues is a podcast that explores bold ideas, community-led research, and inclusive practices shaping the future of healthcare and wellbeing. Post a review and share it! If you enjoyed this episode, please leave us a review. Share it with friends and colleagues. Let’s inspire more people to co-create change. Find out more about the Canada Excellence Research Chair in Health Equity and Community Wellbeing at https://www.torontomu.ca/cerc-health-equity/ Thank you for tuning in!
