Keynote Speaking
Ethical AI in Practice: Balancing People, Models and Systems
Keynote speaker at the 2024 UNC-Charlotte Conference
UNC-Charlotte’s Analytics Frontier Conference 2024
The School of Data Science at UNC Charlotte commits to excellence in education, research, community engagement, and inclusion to shape and lead the future of data science. They teach students to be responsible and ethical data science practitioners, leaders, and researchers in an increasingly data-driven and global society.
Their Needs
To discuss the potential impacts of generative AI for our communities and for industry. Specifically, possible steps to ensure equitable access to the new technologies and ways to protect privacy over the next five years.
Our Solution
Deliver a keynote presentation that noted the current challenges of implementing ethical/responsible AI practices and suggested high priority data governance metrics to help organizations maintain inclusivity
Keynote Description
AI’s impact on everyday life has been particularly disruptive over the last 12-18 months. People within and outside of the AI industry grapple with how to interact with AI. Vetting the accuracy of AI outcomes and assessing the quality of AI tools can be time-consuming and overwhelming even when we have the prerequisite knowledge. For most data leaders, the challenge is compounded by finding practical ways to ensure equitable access to these new technologies and to protect the privacy of our communities.
In this keynote, Dr. Marshall discusses the delicate balancing act among people, models and systems when integrating AI. The notion of fairness, responsible AI framework selection and management are highlighted along with which 7 data governance metrics each organization should prioritize to sustain a more equity-centric digital infrastructure.
The Outcome
Attendees were left enlightened and driven with thought-provoking points to reflect on as the world races to implement AI.
Key Points Discussed:
- Most companies don’t have responsible AI measures in place (roughly 72% have little to no oversight).
- Real risk factors with AI implementation range from revenue and reputational risk to human injustice.
- Data Governance is about governing both data and people (Data quality, availability, incidents…but also data stewardship and education)
- Biases and inaccuracies in source data continue to perpetuate inequitable and/or flawed outcomes.