Pushing the boundaries of knowledge curation
The right combination of these technologies will allow any compatible device to act as an autonomous agent, carrying out intelligent tasks locally, either as an independent entity or as part of a coordinated federated network. As envisioned by Techopedia, a democratized AI system enables:
♦ users to input their data into AI training models without disclosing it to a third party
♦ processing and decision making that operate independent of a centralized authority
♦ developers to distribute their own pre-built training models across a federated network of nodes
♦ more visibility into and transparency over an AI model’s processing activity
This is partly in response to the transition of OpenAI and ChatGPT from a purely open system to a mostly proprietary commercial system. Decentralized AI attempts to move AI development away from such centralized providers toward smaller researchers who innovate as part of a more transparent open source community.
Some are starting to view this as moving from beyond the Internet of Things (IoT) to the Internet of Knowledge (IoK). Those leading the effort include SingularityNET founders Ben Goertzel, of beneficial AI and AGI (artificial general intelligence) fame, and Ethereum co-founder Charles Hoskinson. They and their growing community are forming a one-stop AI business ecosystem: “a central hub for creating, editing, and managing your AI services and the tools to launch those services to a global market.” You can review their strategy for 2024 at singularitynet.io.
And if you’re wondering what role KM might have in all of this ...
At long last, the knowledge library of the future has arrived
Along with universities, libraries have long been recognized as custodians of the world’s knowledge. They go as far back as 4,500 years to the well-organized clay tablets of the Royal Library of the Kingdom of Ebla, located in present-day Syria. Fast-forward to today’s digital realm, and libraries are still at the forefront.
One example is the Harvard University Library. Through its Advancing Open Knowledge Strategy, it aims to democratize knowledge by creating a “global knowledge commons.” This includes integrating the libraries of Harvard’s various schools and departments in true fediverse fashion. Aided by Harvard Law School’s Library Innovation Lab (LIL), planned and ongoing projects cover the entire knowledge lifecycle, from acquisition to preservation to disposition.
On the knowledge acquisition end of the spectrum, LIL has a 3-year program on Democratizing Open Knowledge. One area of particular interest is the Archival Communities at the Edges, which cracks open centralized archives, making its knowledge available to underrepresented communities. Another is the Long Data project, with time capsule encryption, which uses open source software to help make newly released knowledge more durable and robust so it can be rereleased at a specified moment in time, whether that is many decades, centuries, or millennia down the road. No more “Here today, gone tomorrow.”
As KM’ers, we know all too well how our knowledgebases can become cluttered with unwanted “junk.” But one person’s trash might very well be another person’s treasure. LIL’s Trash Exchange is a planned web application for collecting and exchanging digital materials users have deemed worthy of disposal. By allowing users to dig through others’ digital trash, the app can help build a better understanding of the value of discarded artifacts, thereby improving the role of curation.
KM’ers and librarians, unite!
Still looking for that kilobyte needle in the 150 ZB haystack? The computational power of AI and analytics can plow through that staggering volume of content and perform entity-association extraction. Then we KM’ers can work with subject matter experts to validate and make sense of what comes out, along with constructing ontologies to provide enriched meaning for more accurate search and retrieval. Librarians can then apply their skills to categorize and organize the validated outputs so they can be easily navigated, discovered, and applied.
By working closely together, humans and machines will engage in a continuous loop of learning while improving both efficiency and effectiveness. This will promote and enable truly inclusive and transparent democratization of knowledge, creating the possibility for new insights and innovative breakthroughs literally by anyone, anywhere.