Abstract
Some personal artificial intelligence assistants may present a trade-off between collaborative functionality and user data privacy. The disclosed technology describes a decentralized architecture where a personal artificial intelligence agent can operate on a user's local computing device (e.g., a smartphone, smart watch, wearable device, laptop computer, augmented reality glasses, etc.). The agent can utilize a locally trained personal likeness model and a secure on-device data vault to process information. A decentralized collaboration module may enable peer-to-peer communication between agents, governed by user-managed permissions. During collaboration, a privacy-preserving query protocol can be used to exchange synthesized, abstract knowledge summaries rather than the underlying raw data. This architecture can facilitate collaborative knowledge sharing sourced from a user's network while helping to preserve the privacy of the source data and reducing reliance on a central server.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Gupta, Gaurav and Tendulkar, Ashish, "Decentralized Architecture for Collaborative AI Agents Sharing Synthesized Knowledge.", Technical Disclosure Commons, ()
https://www.tdcommons.org/dpubs_series/8821