VOICE PROTOCOL, operated by Japan Realter Co., Ltd., has open-sourced the verification model for its core technology, DEE (Decision Evidence Engine), which aims to make decision-making processes verifiable. DEE is a protocol that records who, what, how, and the conclusion of a decision in a tamper-proof manner, allowing third parties to verify it. 'Proof' in DEE refers to the reproducibility where the same result can be recalculated from the same input. This technology addresses the challenge of 'accountability' in the AI era and is expected to be utilized in corporate governance, financial/credit assessment, AI decision-making rationale verification, and public policy.
VOICE PROTOCOL, operated by Japan Realter Co., Ltd., has open-sourced the verification model for its core technology, DEE (Decision Evidence Engine). DEE is designed to structure decision-making processes to be verifiable post-facto. The initiative addresses the growing issue of 'black-box' AI decisions and the need for objectively verifiable mechanisms beyond mere explanations. DEE records the entire decision-making process (who, what, how, conclusion) in a tamper-proof, third-party verifiable format. 'Proof' in DEE means 'the same result can be recalculated from the same input (event history, snapshot, verification rules),' ensuring reproducibility. The DEE pipeline includes Event, Ledger, Snapshot, Freeze, Bundle, Verify, and Evidence, guaranteeing deterministic recalculation. Unlike traditional logs or reports, DEE enables active verification. It guarantees the reproducibility and verifiability of the decision process, not the correctness of the outcome itself. A public web portal (/verify) allows third-party verification of snapshotHash and reportHash consistency. The reference implementation is open-sourced on GitHub to ensure transparency. Future applications include corporate governance, financial/credit assessment, AI decision verification, and public policy/consensus building.
FACT BOX
- Source: PR TIMES
- Category: 新技術/プロトコル発表, オープンソース化, 製品機能更新
FAQ
What is the primary purpose of the DEE (Decision Evidence Engine) protocol?
The primary purpose of DEE is to structure decision-making processes in a way that makes them verifiable after the fact, addressing the 'black-box' nature of many AI decisions.
How does DEE ensure that decision-making processes are verifiable?
DEE records the entire decision-making process, including who made the decision, what information was used, how the decision was reached, and the final conclusion, in a tamper-proof and third-party verifiable format.
What does 'proof' mean in the context of the DEE protocol?
In DEE, 'proof' signifies that the same result can be recalculated from the same input, such as event history, snapshots, and verification rules, thereby ensuring reproducibility.
What is the difference between DEE's verification and traditional logs or reports?
Unlike traditional logs or reports, DEE enables active verification and guarantees the reproducibility and verifiability of the decision process itself, not necessarily the correctness of the outcome.
What are some potential future applications for the DEE protocol?
Future applications for DEE include corporate governance, financial and credit assessment, verification of AI decisions, and public policy and consensus building.