In an earlier Lucidea series, I set out the Lexicon Framework for Legal KM and explored the reasons many technology initiatives succeed or fail. Implementing artificial intelligence presents a different challenge because tool adoption is rarely the problem.
Unlike many previous technologies, people are already using AI tools enthusiastically. Employees are experimenting with ChatGPT, Claude, Microsoft Copilot, NotebookLM, and a growing ecosystem of AI-powered tools. The challenge is no longer persuading people to engage with the technology, but understanding where human expertise still fits into the picture.
That question becomes more important as AI capabilities continue to expand. In this article, I explore some of the opportunities, risks, assumptions, and possibilities emerging from current discussions around AI and knowledge work, as well as what they might mean for the future role of information professionals.
The Most Interesting AI Projects Begin with Human Curiosity
Every day, there seems to be a new discussion about what AI technology can accomplish when combined with imagination, ingenuity, and creativity. The most interesting AI applications do not simply replace human expertise. Instead, they enable people to ask (and answer) questions that were previously difficult or impossible to explore at scale.
Examples include:
- Historical and archival analysis. Historians and archivists are exploring how AI can reveal patterns across vast collections of texts, helping identify cultural shifts, emerging themes, and previously overlooked relationships.
- Ancient language decipherment. Linguists are applying machine learning techniques to rare and partially understood scripts, using statistical and structural analysis to identify potential connections with known languages.
- Fraud detection and forensic analysis. Organizations are using AI to examine large volumes of financial records, communications, and transactional data to identify anomalies, inconsistencies, and patterns that may indicate fraud or misconduct.
- Scientific discovery assistants. AI-powered research tools, such as the controversial SciBot, are reimagining how researchers interact with the scientific literature, allowing users to interrogate millions of papers through natural language rather than traditional search interfaces.
These breakthroughs are often presented as evidence of AI’s capabilities. Yet, in each case, the breakthrough began with a human question. Someone recognized a challenge that traditional methods struggled to address and used AI to explore it in a new way.
AI accelerated the analysis, but people still framed the problem, interpreted the results, and determined whether the findings were meaningful.
Access Is Not the Same as Understanding
One of the enduring lessons of knowledge management is that information and understanding are two very different things.
Organizations have spent decades capturing knowledge through documents, repositories, procedures, and systems. Yet experienced practitioners know that documentation rarely tells the whole story. It can explain what happened and how something was done, while the reasoning behind those decisions gradually fades from view.
AI excels at processing information. It can summarize documents, generate reports, identify themes, and provide answers in seconds. What it often struggles to provide is the context that gives those answers meaning.
Why was a particular decision made? What alternatives were considered? Which risks were accepted? When should an established process be challenged rather than followed?
These are questions of judgment.
Knowledge transfer has always involved more than moving information from one place to another. It depends on observation, conversation, experience, and the gradual development of expertise. AI may make information easier to access, but true understanding still requires people to engage on a deeper level.
The Value of Strategic Friction
Every generation of information technology promises to make knowledge easier to access. Search engines reduced the need to browse shelves, enterprise search shortened the hunt for documents, and automation streamlined routine processes. AI now promises something even more ambitious: answers without searching.
AI can summarize documents in seconds and surface relevant information almost instantly. Yet recent discussions around AI suggest that efficiency alone may be an incomplete measure of success. When information appears authoritative, people may begin to accept AI-generated answers without fully understanding how they were reached.
Knowledge management has long recognized that understanding develops through engagement rather than passive consumption. People build expertise by questioning information, testing ideas, discussing alternatives, and applying judgment in real-world situations.
This is where the idea of strategic friction becomes useful. In high-stakes environments, a pause for reflection, verification, or professional review is not necessarily an obstacle to progress. It can be the mechanism that improves the quality of decisions.
Libraries, archives, legal information teams, and knowledge managers have been applying this principle for years. Peer review, source evaluation, governance processes, and professional sign-off all introduce moments that require people to slow down, challenge assumptions, and think critically.
The goal of AI should not be to eliminate those moments. It should be to support better decisions by ensuring that human judgment remains part of the process.
AI Can Record the How. People Must Preserve the Why
Perhaps the most important lesson emerging from current discussions is that AI changes the nature of knowledge work rather than eliminating it.
The more capable AI becomes, the more important the human role becomes. As machines take on more of the work of finding, summarizing, and processing information, people become increasingly responsible for setting direction, exercising judgment, and ensuring that technology serves meaningful goals.
Information professionals have operated at this intersection between information and purpose. They’ve spent decades organizing information, creating metadata, building taxonomies, connecting concepts, and preserving institutional knowledge.
The AI revolution is creating new ways to exploit information, while reinforcing the value of human skills: stewardship, interpretation, critical thinking, and judgment.
Context, Stewardship, Critical Thinking, and Accountability
When I began exploring this topic, I was trying to understand where human expertise still fits in a workplace increasingly shaped by AI tools.
What emerged from the research was not a diminished role for information professionals, but a changing one. AI can process information at extraordinary scale, but it still relies on people to frame problems, interpret results, apply judgment, and ensure that knowledge is used responsibly.
As AI becomes a routine part of knowledge work, the role of the information professional moves beyond managing collections, repositories, and systems. Increasingly, it involves providing the context, stewardship, critical thinking, and accountability that AI systems cannot reliably provide for themselves.
AI may be able to document what was done and describe how a process works. People must preserve why decisions were made, determine whether those decisions remain appropriate, and take responsibility for what happens next.
The tools may change. The need for human expertise does not.
Frequently Asked Questions
Why does AI need human judgment?
AI can analyze and generate information, but people must define the purpose of the work, evaluate the reliability and relevance of its outputs, interpret organizational context, and take responsibility for decisions.
What role do information professionals play in AI adoption?
Information professionals help organizations prepare and govern their information, evaluate sources, preserve context, create useful metadata and taxonomies, establish review processes, and ensure that AI-supported decisions remain accountable.
What is strategic friction in AI use?
Strategic friction is a deliberate pause for activities such as source verification, critical review, risk assessment, or professional sign-off. Rather than obstructing progress, it helps prevent inaccurate or poorly understood AI outputs from shaping important decisions.
0 Comments