BFT data consensus algorithm is the key mechanism for PublicAI to guarantee the quality of data which consists of two phases:
**AI Screening Phase**
Once the uploader uploads the data to the platform, our corresponding AI agents will filter it as the first step to mitigate the cost of human verification. Those AI agents are various in terms of data.
**Human Voting Phase**
After the uploaded data is accepted by AI agents, it will enter the human verification phase. Consensus is the same number of votes in the same direction. Data sample quality is binary: if a data sample is partially good then it's considered overall flawed.
We have three roles in the protocol to verify: **scouts, guards, and judges** with three hyper-parameters for them respectively as consensus thresholds.
Concensus Reach Scenarios
Take an example from the above graph, once a data sample is accepted by the data consensus algorithm there should be at least either 50 votes from scouts, 5 votes from guards, or 1 vote from a judge.
---
# Agent Instructions: Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.
Perform an HTTP GET request on the current page URL with the `ask` query parameter:
```
GET https://docs.publicai.io/publicai-documentation/usdpublic-tokenomics/bft-data-consensus-algorithm.md?ask=
```
The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.
Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.