Once storage is that enormous, the only way to navigate data and information is through AI-based technologies, particularly machine learning. Advances in AI have added significantly to people's ability to discover insights previously hidden. Pattern recognition, for one, triggers an under- standing that couldn’t be gleaned from a manual analytical approach.
Artificial/augmented intelligence
A groundswell of interest in generative AI was triggered by the arrival of OpenAI’ s ChatGPT as a widely accessible tool. Within a few weeks, millions of people signed on to create essays, research papers, poetry, articles, job applications, memos, computer code, jokes, short stories, and more. As with most large language models, its original training set is proprietary information. But since ChatGPT is constantly learning from the questions it is asked and the prompts it is given, the training set grows exponentially on a daily basis. The trick with opening up tools like this to the public is the possibility that they will be manipulated for uses other than the intended ones. The introduction of biased, false, and misleading information into the algorithms is just too easy in a publicly available tool.
ChatGPT is not the only large language model that has potential for improving KM. Google’s BERT (Bidirectional Encoder Representations from Transformers) has been used for several years to analyze sentence structure, determine contextual meaning, and improve relevancy in search. Since Google made BERT open source, it can be used by many companies. Google’ s LaMDA is another stab at a large language model that might replace, or at least augment, web searching.
For KM purposes, internal search using machine learning, natural language processing, and other AI-based technologies has a bright future. The ability to suss out intent from search queries is enormously helpful in guiding people to correct answers and the information they need.
Excitement about new technologies and their applicability to KM and the wider world should not obscure a basic fact: At its core, KM is about people. It’s people who create, curate, evaluate, and share knowledge. They are the ones who worry about securing knowledge, making sure that information is protected and does not fall into the wrong hands. Maintaining customer and employee privacy, assuring that laws and regulations regarding data are adhered to, and keeping sensitive information away from competitors is both a technical and human task.
Knowledge sharing comes naturally to humans. Marc Vontobel, CEO of Starmind, posted recently to LinkedIn that he hated the term “knowledge management” because it connoted forcing people to do things they didn’t want to do. He went on to say that with an AI-supported approach, people are happy to help others by sharing knowledge, since the AI takes over the laborious, time-consuming part of the job.
Pairing human knowledge with technologies that allow for data extraction, information analysis, and knowledge insights is the future of KM.