1 Executive Summary
Innovative and dynamic technologies require enterprises to be flexible to stay current. Distributed and decentralized artificial intelligence is one of these fields. When defined very generally, distributed or decentralized AI (DAI) refers to spreading ownership, responsibility, or tasks across many parties, which can produce heightened results, whether it be in accuracy, robustness, customization, or geographical spread. Closer inspection of DAI market products shows that there is little agreement from vendors on what the terms really mean and what the underlying issue is that the solution should actually address.
The terminology is undergoing constant refinement, which makes misinterpretation easy. Decentralized AI (DAI) is the term with the most traction today, capitalizing on the blockchain hype and evokes images of platform and networking effects. Distributed AI – still sometimes referred to as decentralized AI – is a more traditional term that is used to describe the algorithmic capabilities of multiple AI to communicate with each other to solve a common problem. These agents may be spread geographically, have heterogeneous functions, or both, making the task of coordinating their efforts a R&D category all on its own.
This is not simply a question of semantics; these different definitions yield different use cases for the enterprise, which range from secure data pooling, autonomous vehicle fleet management, development of AI marketplaces, and the ability to build and use modular AI services.
Both conceptualizations of DAI can lead to positive value gains for an enterprise but requires any interested enterprise to have a nuanced understanding of what they wish to gain from integrating such technology. This leadership brief serves as an introduction to the different value propositions that DAI vendors present, and equip the enterprise with the right questions to ask of themselves and potential vendors to determine if there is an alignment of goals.