Mastercard, eBay, and Capital One recently participated in the Women in AI Breakfast, sponsored by Capital One for the third consecutive year. The event, held as part of VB Transform, focused on the emerging field of generative AI and its potential impact on various industries.
The breakfast discussion delved into the challenges and opportunities presented by generative AI. Emily Roberts, SVP and Head of Enterprise Consumer Product at Capital One, highlighted the need for organizations to build continuously learning structures to fully leverage the potential of this technology. She emphasized the importance of diversity and representation in the development and implementation of generative AI, as it offers a greater chance to establish equity as the foundation.
Data remains a critical factor in the development of generative AI, particularly in relation to public large language models (LLMs). JoAnn Stonier, Fellow of Data and AI at Mastercard, acknowledged the inherent challenges associated with historical data that might contain biases or inequities. She emphasized the need to define boundaries and expectations for outcomes, as well as to proactively address potential issues, especially within sectors like finance where fraud prevention is vital.
Xiaodi Zhang, VP of Seller Experience at eBay, stressed the importance of establishing guardrails and constraints to ensure equitable and unbiased results. She emphasized the need for a continuous learning approach, experimentation, and flexibility within the industry.
While companies are cautious about launching new use cases due to the remaining risks, they are investing in internal innovation to explore the possibilities of generative AI. For instance, eBay recently conducted a hackathon focused entirely on gen AI to leverage the capabilities and creativity of their employees. However, careful thought and consideration are paramount, as the exploration of generative AI must be done in a thoughtful and responsible manner to eliminate bias.
Mastercard encourages internal innovation but recognizes the need for proper guardrails and frameworks to evaluate and submit use cases. They are exploring applications in areas such as knowledge management, customer service, advertising and marketing services, and interactive tools for customers. However, they aim to eliminate biases before making these applications public, especially in critical areas like healthcare and legal decision-making.
Regulations have begun incorporating generative AI, and companies are working to understand the necessary documentation and reporting requirements. They will need to demonstrate the thoughtfulness and refinement of their projects to align with regulatory expectations.
Capital One has already rebuilt its fraud platform using cloud technology, data, and machine learning. They emphasize the importance of experimenting, testing, and learning in a well-managed and transparent manner, especially as regulations and standards evolve.
Overall, the event highlighted the need for organizations to embrace generative AI cautiously, while investing in internal innovation, establishing guardrails and inclusive practices, and adapting to regulatory decisions. The industry acknowledges the power of generative AI but continues to prioritize responsible and ethical implementation to ensure equitable and unbiased outcomes.
