The future of KM is not simply AI

It’s tempting—oh so tempting—to extrapolate from today’s hot-button topic of AI to assert that AI represents the complete future of knowledge management and knowledge sharing. And not just AI but, specifically, generative AI (GenAI). Even more specifically, ChatGPT. In the future, that knowledge management will be indistinguishable from ChatGPT seems to be an underlying theme in many prognostications. Those prognostications are very likely to be proven incorrect.

Equating GenAI, whether in the form of ChatGPT, Claude, or another chatbot, to knowledge management does a disservice to the other aspects of KM and puts way too much emphasis on GenAI and what the company OpenAI is developing. OpenAI, remember, is far from the only player in the GenAI space, although it’s getting most of the attention right now. Every company under the sun seems to be jumping on the AI bandwagon and proclaiming that their products and services are now AI-empowered, AI-driven, AI-augmented, AI-compatible, and, well, AI-everything. The interest in AI technologies is unprecedented; the hype is unrelenting.

The danger of prognostication based on a single datapoint is the assumption that the situation will advance in a straight line going forward. In the real world, this is extremely unlikely. Unknown factors can, and probably will, transform that straight line into a jagged curve of some sort. It’s always problematic to assume simple continuity and linear progress. Other technologies, failures of implementations, and “out-of-the-blue” developments will affect how KM is practiced in the next year or two.

In the near future, we will see how many of the promises being made around AI will actually be realized. It is inevitable that not everything will prove successful. There will be a “shakeout” period when some of the products and services announced will fade from view. AI will most definitely, however, play a role in the future of KM. GenAI’s ability to create new knowledge is just short of magic. Yet the creation needs to be balanced with human understanding and critical thinking. Just because GenAI can present us with something that sounds plausible doesn’t mean it’s correct. Validity and accuracy are too often lacking in GenAI responses unless serious oversight is in place.

This is of particular concern within the enterprise. No one wants incorrect information to be disseminated either within or outside a company. A plausible-sounding policy that contains an error can create havoc, particularly if that policy concerns a compliance issue or the protection of trade secrets. A customer receiving inaccurate information from an agent can have reactions ranging from being mildly disgruntled to screaming angrily to threatening lawsuits. None of these are good outcomes. Looking ahead, they will be fixed, or at least everyone hopes they will be fixed, as the technology matures and people become more adept at using the technology.