To prevent AI from running out of control, we should begin by ensuring that humans and AI share a common understanding. At the core of this is conceptualization technology. In other words, both sides need a shared language that makes it possible to understand “what kind of judgment should be made under what kind of pattern.”
For the past 25 years, Mindware Research Institute has been engaged in SOM-based data mining, while researching the integration of traditional epistemological philosophy concerning “concepts” with machine-learning approaches. We are not a startup simply jumping on the recent AI boom.
AI has learned from vast amounts of information across the internet, which means that much of what people used to search for online can now be answered simply by asking AI. What will this lead to? In the realm of information, differences between individuals and between companies will diminish. As long as people make rational decisions, the probability that everyone arrives at the same conclusion will increase. This opens the door to an existential crisis: people may believe they are making their own decisions, while in reality they are being guided entirely by AI.
In this AI era, it is becoming increasingly important for both individuals and companies to have their own distinctive standards of judgment — standards that reflect who they are and what their organization stands for. We continue to think seriously about how to make AI act according to our own criteria, rather than allowing it to make arbitrary judgments on our behalf.
To implement our technology in society, we have decided to begin with the construction and operation of knowledge bases. Seen from the broader vision described above, this is a case of thinking big while starting small. Even in knowledge bases, however, the value of conceptualization is enormous. Conventional RAG has limitations precisely because it lacks this process of conceptualization.
Please try our Concierge Builder. In just a few minutes, you can create a response system that works reliably. We have already received comments such as, “It feels almost unbelievable that something we struggled with so much before can now be done this easily.”
