Synopsis: Dr. Arthur O’Connor, academic director of data science for the School of Professional Studies at the City University of New York (CUNY) and author of "Organizing for Generative AI and the Productivity Revolution", delves into how advanced reasoning capabilities being added to large language models (LLMs) will profoundly alter business processes as costs continue to rapidly decline.
In this Techstrong AI interview, Mike Vizard speaks with Dr. Arthur O’Connor, academic director for data science at the City University of New York, about the evolving landscape of AI model development, focusing on the recent buzz around DeepSeek. O’Connor explains that DeepSeek’s success shows it’s not just about building larger models with massive datasets and expensive GPUs—it’s about creativity, smarter data design, and high-quality targeted training. He points out that distilled models, which are optimized for specific tasks, are gaining traction because they are more efficient, cost-effective, and practical for real-world business applications compared to massive, general-purpose models.
The discussion then shifts to the organizational challenges businesses face when adopting generative AI. O’Connor stresses that many companies are still structured around older digital models, with data science expertise siloed away from business operations. To successfully integrate AI, companies must decentralize data science knowledge while centralizing governance, standards, and security policies. He warns that without careful oversight, organizations risk training models on proprietary information and lacking the monitoring needed to detect model drift. The traditional deterministic workflows in businesses also clash with the probabilistic nature of generative AI, requiring a major rethink of business processes and performance evaluation.
Finally, O’Connor discusses the broader cultural and workplace impacts of AI. He suggests that workers are evolving from knowledge creators to knowledge supervisors, curating and overseeing AI outputs rather than producing everything manually. However, he cautions that over-reliance on AI can lead to cognitive offloading, diminishing critical thinking skills over time. To navigate this new era, he advises organizations to stay flexible, focus on fundamentals that don’t change—like customer needs—and invest heavily in upskilling employees. His recent book, Organizing for the New Productivity Revolution, provides further guidance on adapting to this technological shift.
