Data governance is in greater demand than ever, considering how the importance of data quality, integrity, and relevance continues to grow in our AI-centric world.
As organizations explore and adopt generative AI and other sophisticated data-infused applications, they will need to rethink how they organize data governance practices and align those with their machine learning, application development, and IT systems management operations. That’s because generative AI presents new challenges in data trustworthiness, transparency, ethics, equity, and safety.
In his keynote, James Kobielus, TDWI’s senior research director for data management, will discuss the new shape that data governance is assuming in modern organizations. He will highlight the following key aspects of this evolved practice:
- Centralize governance of generative AI data, models, and other assets used in the building, training, and operationalization of these applications
- Build transparency into the AI-driven generative models used in the data engineering pipeline
- Implement strong governance and auditability over training data used in generative AI operations
- Incorporate data privacy, bias, transparency, toxicity, and safety specialists, roles, and controls to the governance process
- Institute close curation over all AI-generated output to ensure its veracity, fairness, freedom from toxicity and bias, and other aspects of fitness for downstream uses
- Establish a program of appropriate compensation for providers of source content that is used to train AI-generated content engines deployed by the enterprise