The CI Sector’s publication on "Challenging systematic prejudices: an investigation into bias against women and girls in large language models (LLM, 2024), highlights systematic LLM biases and offers recommendations for policy and practice to support women's inclusion in the digital age.
Activity Type Research/Reports/AssessmentsPolicy/Regulatory Guidance
UNESCO’s Red Teaming AI for Social Good Playbook (2025) is a practical guide to test AI systems for gender bias and technology-facilitated gender-based violence (TFGBV), empowering organizations and communities with tools to test AI for harmful biases—especially those impacting women and girls. Developed in collaboration with Humane Intelligence, the Playbook offers a practical, step-by-step guide to help users design and implement their own Red Teaming events and includes concrete examples that reveal gender bias in AI outputs. The exercises provided in the Playbook allow participants to test AI systems in a controlled environment by simulating real-world prompts and analyzing responses. No technical background is required, making the Playbook accessible to educators, policy experts, civil society, and others investigating the ethics of AI and its societal impact.
An AI capacity building project that strengthens AI/ML capacities in developing countries through mentoring activities and workshops on tinyML technology. The goal is to create a sustainable ecosystem where local developers can innovate with AI/ML for development and prosperity.
Activity Type Networks/Mentorship/ExchangeThe project aims to train 200 independent journalists and community media in Paraguay on reporting on technology, including on Artificial Intelligence and its ethical use. The goal is to promote the production of high-quality journalistic content while strengthening the involvement of communities in debates around technology and artificial intelligence.
Activity Type Awareness/Advocacy