By Katrina Pugh, Ph.D., Lecturer in the Information & Knowledge Strategy (IKNS) Program at Columbia SPS
AI brings stunningly fast answers by accessing and processing large amounts of the world’s knowledge. AI produces answers with seemingly humanlike awareness and has the potential to compress cycle times and to elevate all of us—new learners and sage experts, alike. However, the use of AI in knowledge work poses risks: inaccuracy, manipulation, compromised learning, and social isolation.
To address these challenges, a group of colleagues and I started wondering whether strategies borrowed from knowledge management (KM)—one of the root disciplines of the IKNS curriculum—could counter these shortcomings and offer an outline for human-AI partnership backed by information science.
In our article “Knowledge Managers Bring Collectivity, Nostalgia, and Selectivity to the AI Ecosystem,” I am joined by Andrew Trickett, Jonathan Ralton, Marc Solomon, and Eve Porter Zuckerman in presenting three case studies: one from a global architecture, engineering, and construction services firm; one from a tech developer; and the last from a large multiline insurer. Through these studies, we propose a new type of human-AI system that applies artificial intelligence in tandem with human intelligence, using the KM skills of collective sense-making (which we call “collectivity”), curation of exemplars (“nostalgia”), and directed, out-of-box thinking (“selectivity”).
These cases illustrate three important ways that knowledge managers scrutinize the product of AI, inform the training of AI, and marry AI’s scale and speed with our human ingenuity. They also show how knowledge management combats many of AI’s risks by using collectivity, nostalgia, and selectivity.
Our first case study looks at the use of collective engagement, or “collectivity,” within a London-based multinational professional services firm. Before the firm implemented AI, a community forum had been established to curate lessons learned and collect comments about client requests and past practices. After the firm introduced AI, this practice continued and was expanded, allowing knowledge-holders within the company to design prompts for AI and co-edit them, curating a collection of AI insights and prompt engineering methods for the community to access, refer to, and learn from.
Large language models (LLMs), which can be used to retrieve information and generate text, often have different priorities than humans, and may, for example, respond to a prompt with the most recent response rather than the most fitting one. To counter this, the LLM can be trained on a continuous basis with human-vetted “exemplars” (artifacts selected for specific parameters). As a result, the AI’s scoring of works-in-progress can be more accurate, comprehensible, and scalable. The second case study applies this KM approach that we call “nostalgia”––curation of historical content based on specific features––at a U.S.-based technology development company. The nostalgia approach provides a blueprint for training the AI continuously on historical exemplars that have notable strengths but would have been overlooked by AI alone because of their age or their having other unrelated features. The team proposed a process that would train AI on these exemplars, then use AI to score current work products.
Another flaw of LLMs is that they are usually trained to recognize and produce content that is similar to a prompt, focusing on convergence, and the tried-and-true over the novel. We note in the article that “Undifferentiated LLMs fail to apply a selective focus, missing novel opportunities and even competitive advantages.” In our final case study, AI was used at a U.S.-based multiline insurance company to pull back peer organizations’ unique ESG reporting approaches, calculating each of their year-over-year commitment-fulfillment scores. Compiling these data into a shared taxonomy, or “master data,” translates or normalizes the peers’ categories. With selectivity, it is possible to benchmark the insurance organization against its peers quickly and with more innovative measures of ESG.
A reader might think that we’re disparaging AI, as it may dampen creativity. We note that “AI-powered creative acts [e.g., art, innovation, research] tend to short-circuit our self-expression and devalue or preempt human-human interaction.” On the contrary, we find that knowledge managers like those in the SIKM Boston community, and in the IKNS program, can successfully craft a partnership with AI.
Collectivity, nostalgia, and selectivity are foundational KM capabilities for using AI, which keep accuracy, productivity, and human agency intact. We believe that when KM'ers build cross-organizational, cross-functional, cross-temporal or cross-disciplinary bridges, we overcome many of the accuracy, efficiency, and relevancy limitations of AI. These bridges also make us all more resilient and discerning humans: AI cannot substitute for our humanness, but it can bring us scale, reach, and speed.
Practitioners in knowledge-centric jobs can get the AI-human combination right by deliberately investing in these bridges. They and their organizations need to come together to scrutinize AI’s results (what we refer to as “collectivity”), to tap past content and experts to construct exemplars (“nostalgia”), and to be inspired and educated by outside approaches (“selectivity”). The onus is on us to reimagine our roles as orchestrators of a future where we collaborate better not only with AI, but with ourselves.
About the Program
The Columbia University M.S. in Information & Knowledge Strategy (IKNS) degree provides students with foundations in information science, organizational psychology, and change management as well as practical skills in project management and executive leadership.
The STEM-accredited program is available full-time or part-time, online or in-person, on Columbia’s landmarked campus right here in New York City. To maximize opportunities for networking and community building, our online students join our New York-based students on Columbia’s campus for three in-person residencies during their studies.
Students train under world-class faculty, including former and current executives from Google, IBM, NASA, and Oliver Wyman, and join a powerful global alumni network in coveted positions, including at Alphabet, Goldman Sachs, Nike, Pfizer, and the World Bank.
For more IKNS insights, news, and events, please go to our website, connect with us on LinkedIn, or attend one of our online info sessions. Visit the School of Professional Studies website to learn more about the SPS student experience.
The final fall 2025 application deadline for the IKNS program is June 1. Learn more here.
