Toju Duke: Redefining Ethical AI & Inclusive Innovation

Transitioning to Ethical AI Leadership

• 🚀 Toju Duke shares her journey from Google to founding Bedrock AI and Diverse AI, driven by a desire to work on critical issues without corporate bureaucracy.
• ✨ She found the transition highly rewarding, enabling her to pursue her vision for responsible AI more directly.

Developing Responsible AI Frameworks

• 💡 At Google, Toju worked on large-scale models like Lambda (powering Gemini) and identified associated risks and harms.
• 🛠️ She built processes for due diligence and ensuring transparency through data cards to assess risks before model launches.
• 🎯 Bedrock AI was launched to bridge the gap between understanding AI problems and providing practical AI ethics solutions, including risk assessments and AI principles for organizations.

Fostering Diversity in AI

• 🌍 Diverse AI aims to make the AI field more diverse and integrate diversity directly into AI technology.
• 🧠 It addresses bias and discrimination in AI models by focusing on improving training data sets and diversifying the teams that build these systems.
• 🖼️ A key project involves enhancing text-to-image generators by increasing diversity in their training data sets, including images and captions, with community feedback and contributions.

Critical Considerations for AI's Future

• ⚠️ Toju urges global leaders to slow down the pace of AI development to combat misinformation, hype, and the resulting erosion of trust in AI systems.
• ✅ She emphasizes the importance of developing a clear AI strategy and educating the public on both the benefits and risks of AI.
• 📈 Leaders must focus on reskilling and upskilling employees to adapt to job changes brought by AI, preparing the workforce for new opportunities.
• ⚡ A crucial, often overlooked concern is the energy consumption of large language models, advocating for intentional and deliberate use of prompts to mitigate environmental impact.