Knowledge management in 2021: KM enters a new stage

This year, KMWorld went virtual with 4 days of live keynotes and conference presentations delivered to attendees' screens.

What dominated KMWorld Connect 2020 was the emphasis on knowledge graphs. A good way to put this in perspective is to say that we have arrived at the fourth stage of KM, the “knowledge graph stage." Probably “ontology stage" would be a more accurate name, as well as broader and more inclusive, but knowledge graph stage is crisper, more memorable, and will probably stick. It has for some years been recognized that KM developed through three clearly recognizable stages.

KM stage 1: Information technology

The first stage of KM was driven primarily by IT, information technology, specifically the internet, and the realization that Intranets were a marvelous tool for sharing information around geographically dispersed organizations. The salient point is that the first stage of KM was about how to deploy that new technology to accomplish more effective use of information and knowledge.

The slogan of the period, "If only Texas Instruments knew what Texas Instruments knew," captures that spirit. The illustrative hallmark phrase of Stage 1 was “best practices,” later to be replaced by the more politic “lessons learned,” recognizing that what might be appropriate in Singapore might not fit so well in Stockholm.

KM stage 2: Human aspects and corporate culture

The second stage of KM emerged when it became apparent that simply deploying the technology and making that valuable information available was not sufficient to effectively promote and accomplish information and knowledge sharing. Human and cultural dimensions needed to be addressed. The second stage might be described as the "If you build it, they will come is not an effective strategy" stage. In fact, it can too often lead to quick and embarrassing failure. Most consequentially, it became clear that KM implementation would involve changes in the corporate culture, in many cases rather significant changes, even up to and including the organization’s management structure.

KM stage 3: Taxonomy and content management

The third stage was about the organization of information, the taxonomy stage.  It developed from the awareness of the importance of content and, in particular, the awareness of the importance of the ability to retrieve that content and, therefore, of the importance of the arrangement, description, and structure of that content. Following on from another good alternative description for the second stage that “KM is no good if they don’t use it,” the third stage can be described as “KM is of no use if they try to use it but can’t find it.” Way back at KMWorld 2000 a track on content management appeared for the first time, and by the 2001 KMWorld conference, content management had become the dominant track. In 2006, KMWorld added a 2-day workshop entitled Taxonomy Boot Camp, which still thrives today in an expanded form. The hallmark terms for the third stage of KM are taxonomy and (enterprise) content management.

The first three stages of knowledge management evolved fairly rapidly, and KM may be said to have emerged in 1993 with the first conference about KM, organized by Ernst & Young in Boston. By 2006, with the establishment of Taxonomy Boot Camp at KMWorld, the third stage was firmly established. During the next dozen or so years, KM matured and grew. Some aspects such as capturing the expertise of emeritus and about-to-be-emeritus employees became more important but on the whole there were no major shifts in the direction or understanding of KM. The emerging knowledge graph stage is, however, a major change.  

KM stage 4: Knowledge graphs 

The knowledge graph stage represents the awareness that relating terms such as parent, child, sibling, equivalent, etc., are not sufficient to adequately explain relationships. To explain relationships between things, you need the predicate, as in professors (thing, subject), teach (predicate), students (thing, object). You need those triples—subject, predicate object—developed in RDF (resource description framework) and OWL (web ontology language; yes, web ontology language doesn’t seem to map well onto OWL, but the developers just thought that "owl" sounded so much more attractive than WOL). The knowledge graph is the structure or the map (graph) to tie those triples together.

Mentions of knowledge graphs began to appear a few years ago and, by 2019, it could be described as one of the major themes of KMWorld. However, in 2020 it dominated as the major theme.

In addition, this year there was an increased mention of AI. That development was very much related to fourth stage KM. The context was most often the hope that AI and improved KM could get the user, most often the customer and not the employee, expeditiously to the requested or needed information.

We all know the problem and the frustration of responding to a “contact us” prompt at a website, hoping to be able to get in touch with a real person, only to be presented with a canned list of FAQs, none of which are remotely related to one’s question or information need.

The hope and the intent is that developments in AI and the emerging knowledge management graph stage KM will greatly ameliorate that problem. 

Another topic that surfaced for discussion at KMWorld Connect 2020 was Generative Pre-trained Transformer 3 (GPT-3), a language generator with a degree of comprehension that was just released in the summer of 2020. GPT-3 is far more than a language generator that can write prose in the style of Ernest Hemingway or Patrick O’Brian.  It can write code, at least experimentally.  After seeing a number of applications and the code for them, GPT-3 was able to write code for an application based on a simple description of what the application was intended to do. “Keep your eyes open for further developments” was the typical context for the mentions at KMWorld.  Again, the hope and expectation is that linking GPT-3 capabilities with KM system knowledge will provide better answers to information needs.

Finally, one could observe in KMWorld Connect 2020 the continuing change in emphasis toward KM serving the external community of the organization, not so much the internal community.  It is no longer "If only Texas Instruments knew ..."

Knowledge management continues to evolve

In summary, perhaps the best description of KMWorld Connect 2020 is that it is the place where the fourth stage of KM, the knowledge graph stage emerged in full. Will next year be the GPT-3 stage?

Access to all KMWorld Connect 2020 presentations continues to be available to registered conference attendees at https://pheedloop.com/kmwconnect/virtual/?page=lobby.

In addition, if you missed the live event, you can see all the keynotes and conference sessions with the purchase of a replay pass at https://pheedloop.com/kmwconnect/site/register.

Be sure to save the date for KMWorld 2021. The live event will take place November 15–18, 2021 at the JW Marriott in Washington, DC.

References for the development and the stages of KM:

Koenig, Michael E. D., (2000), The Evolution of Knowledge Management, in T. K. Srikantaiah and M.E.D. Koenig, Knowledge Management for the Information Professional, (pp. 23-36), Medford NJ, Information Today for The American Society for Information Science.

Koenig, Michael E. D., and Neveroski, Kenneth, (2008), The Origins and Development of Knowledge Management, Journal of Information and Knowledge Management, 7(4), 243-254.

Prusak, Larry, (1999), Where Did Knowledge Management Come From? Knowledge Directions, 1(1), 90-96.