How To Build Successful AI Models

The panel discussion on building successful AI models featured speakers from diverse backgrounds who emphasized the importance of measurable AI, data ownership, talent acquisition, and ethical considerations in leveraging AI for societal benefits. Key themes included the need for decentralized data, an AI-first platform, patient-driven governance in healthcare, and transparent and accountable AI systems to ensure trust and positive societal outcomes.

The panel discussion on building successful AI models featured speakers representing various industries and initiatives. Mark Greaves, working on the AI 2050 program, highlighted the importance of getting the smartest minds to address key challenges. The program aims to inspire innovation by focusing on the benefits of AI and addressing technical and societal challenges of AI deployment globally. He emphasized the need for measurable AI and managing the socio-economic impact of AI on society to ensure widespread benefits.

Anda Kazus, co-founder of Vanana, stressed the significance of data ownership in AI models. She highlighted the need for high-quality decentralized data to improve AI models and ensure diverse perspectives are represented. Kazus discussed applying decentralization principles from cryptocurrency to AI and data to empower individuals to contribute data in a privacy-preserving way while enhancing data quality and representation.

Rzan Calan, Chief Digital and Payments Officer at TD Bank, shared insights on leading the digital transformation within a regulated bank. He emphasized the importance of talent acquisition, agile operating models, and innovative architectures to infuse AI across all aspects of banking operations. Calan emphasized the need for an AI-first platform that integrates data, compute, models, and applications to deliver personalized customer experiences while maintaining trust and enhancing data security.

Stuart Davis from the South Australian government discussed the initiative to transition healthcare systems towards prevention and wellness using AI. He highlighted the unique data assets held by the state and the importance of patient-driven governance and ethics in leveraging data for healthcare transformation. Davis emphasized the need for trusted execution environments to ensure data privacy and responsible sharing of insights for disease understanding and drug development.

The panel concluded with reflections on overcoming key impediments in unleashing the impact of AI on society. Mark Greaves highlighted the need for calibrated and measurable AI to build trust and manage societal transitions due to AI’s impact on employment and relationships. The speakers emphasized the importance of holistic AI ecosystems, data ownership, and building trust through transparent and accountable AI systems to achieve widespread benefits and positive societal outcomes from AI advancements.