Published on February 5, 2024, 2:28 pm

Generative AI, also known as genAI, has gained significant attention from enterprise CEOs and boards of directors since the release of ChatGPT in November 2022. A report by PwC revealed that 84% of CIOs are expecting genAI to support a new business model in 2024. Undoubtedly, genAI is a transformative technology. However, it’s essential to recognize that it is just one type of AI and may not be suitable for all use cases.

The definition of AI has evolved over time. For example, what was considered AI fifty years ago, such as a tic-tac-toe-playing program, is no longer seen as groundbreaking today. The history of AI can generally be categorized into three different segments.

According to Thomas Robinson, the COO at Domino, generative AI accounts for approximately 15% of use cases and models among chief data and artificial intelligence officers (CAIOs). Predictive AI still remains prevalent in model-driven businesses, and future models are expected to combine predictive and generative AI.

Interestingly, there are already instances where predictive and generative AI complement each other. These include analyzing radiology images to generate reports on preliminary diagnoses or mining stock data to predict future increases in value. This convergence suggests that organizations will require a common platform for developing complete AI solutions.

Developing and deploying complete AI does not necessitate treating each type of AI as a distinct entity with its own stack. While genAI may require additional computing power and network enhancements for optimal performance in certain areas, building an entirely new infrastructure from scratch is typically unnecessary unless dealing with massive scale deployments like Meta or Microsoft.

Similarly, processes for governance and testing do not need to be completely reinvented when incorporating genAI. For instance, mortgage risk models powered by predictive AI undergo rigorous testing, validation, and monitoring – similar practices required by large language models (LLMs) used in genAI. Although there are differences, such as genAI’s tendency to generate “hallucinations,” the overall risk management processes are generally similar.

Domino’s Enterprise AI platform is trusted by one in five Fortune 100 companies to manage AI tools, data, training, and deployment. This platform enables AI and MLOps teams to oversee complete AI development, deployment, and management from a central control center. By unifying MLOps under a single platform, organizations can effectively leverage the benefits of both predictive and generative AI.

To learn more about how to maximize the rewards while managing the risks associated with genAI projects, Domino offers a free whitepaper on responsible genAI. This resource provides valuable insights into harnessing the full potential of genAI within an organization while ensuring responsible implementation and usage practices


Comments are closed.