Decentralized AI Platform for Collective Governance

Consensus AI is the open-sourced, decentralized platform, powered by native cryptocurrency, and built with the vision to improve the governance of private and public organizations — companies, regions, cities, or countries — using artificial intelligence and decentralized services.

Consensus platform will help govern communities, societies and countries in a more cost efficient, transparent and progressive manner. As part of the roadmap to reach this goal we will build digital communities on our network (for a city it will be the community of citizens, for a private company — the community of clients or loyal customers) and give them an opportunity to influence the decisions made by the city, region or their favourite brand.

During the initial project stage we will build and launch the Sentient network, a distributed ledger system capable of performing decentralized machine learning, powered by its own currency, smart contracts and the implementation of trusted nodes.

The two main features of Sentient are:

1. Verification of nodes and smart contract transactions with eIDs, which allows for the creation of governance applications and faster adoption by public sector as a legitimate platform for data processing and decision making.
2. Decentralized machine learning, allowing to keep private or sensitive data with their owners. Using the federated learning approach we will bring the models to where the data stores are and train them locally on the nodes, without the need to upload user data to a centralized server.

Further product direction is divided into 4 phases:

Phase One — Communities

Launch decentralized communities on the network.

Members will be verified through eIDs and will be able to join the communities of relevant cities and countries. For example, a resident of Toronto will be able to join the communities of the district where he or she lives, the community of the City of Toronto and the community of Canada. But they will not be able to join the community of Vancouver at the same time.

Community members will be able to:

• Put forward the proposals to region and city governments
• Vote for existing proposals
• Give feedback to initiatives, submitted by the government of a region/city.

All the voting and polling results will be recorded on the blockchain and verified by eIDs, so they can’t be changed or manipulated.

The initial release of the platform will be developed with the support of Estonian e-Residency and launched in Estonia, and later in other countries as their eIDs become available to join the network. For example, the European eIDAS directive will require mandatory acceptance of each others’ eIDs by all member states by September 2018, signalling electronic ID maturity, therefore making it possible for us to integrate all the European eIDs into the Consensus platform. We expect most other countries to introduce electronic IDs in the next 10–12 years.

Phase Two — Transparent Workflow

Phase Two is to launch the system for the government officials to involve communities in the work on different proposals, tasks and projects — from building a new hospital to organization of a city holiday.

There are similarities to OpenGov features in some aspects, but additionally the community and external experts opinion data will be included in the decision making process and execution — whether to accept or reject the project, include it in the budget etc. Communities will also be able to influence, where applicable, the choice of a contractor for projects, track the execution and give feedback on results. The entire process from an idea to the final project implementation stage will be recorded on the blockchain, which will provide the transparency of decision making and funds distribution (allowing for the allocation of tax budget to be traced and visible to taxpayers).

Phase Three — Data Acquisition

At this stage we will start collecting relevant data from various sources, bringing it into a structured format using predefined templates and semantic categorization, and matching it with objects, events and persons utilizing various meaning extraction approaches. For example, the data can be the salary distribution in a certain city, analysis of the population dynamics, traffic load in various districts of the city, number of requests for medical assistance per year, water and electricity consumption etc. Consensus Foundation will collect this data from open sources, buy from third-party suppliers and offer the network users the option to provide data in exchange for SEN. Next, we will fuse it with data received from our initial communities — what decisions they made for each type of questions, what consequences it led to, how people from every area, city or region prefer to vote on various types of issues.

This will help the engine draw non-obvious conclusions, find hidden patterns and insights. While working on inquiries, officials and citizens will be able to address a large data corpus, accumulated from their city and other cities’ experience, see graphs, stats, forecasts etc.—which will help make and implement accurate, evidence-based, data-driven decisions.

Phase Four — AI Advisor

To be able to propose optimal solutions Consensus AI engine will be trained using the available data. Instead of researching the optimal course of action manually by studying vast amounts of data we envision researchers to be able to make requests to the Consensus AI in a way consumers use Apple Siri or Amazon Alexa (it’s a simplified analogy, but helps illustrate the point). For each task or proposal the Consensus AI engine will take into account the public opinion and other relevant data, simulate all potential positive and negative outcomes, search for optimal decisions based on available data and other cities/regions’ experience.

The artificial intelligence engine we envision will be decentralized and powered by the Sentient network. In addition to the data contribution (optional) each network node will be able to provide a portion of its hardware resources for the AI engine computations, and will be rewarded in SEN for it—each node will be able to determine what percentage of its resources to make available for AI engine computations. As a result the Consensus AI will be powered by the fog supercomputer where nodes will receive rewards in SEN that correspond to the price of compute power.

The Future

We estimate that 2–5% of the citizens are being active in the public decision making and will regularly vote and make their own proposals. Therefore, in the long term, we plan to train the AI engine to predict the public opinion of each community, analyzing information about them in the background.

We envision that in the future Consensus AI will represent the interests of residents and give the officials an immediate feedback on any issue, without distracting people from their usual affairs. People will live their normal lives and provide the AI engine with their data for a reward in SEN. Using our approach, the engine will bring the models to where the data stores are and train them on the encrypted user data, so that citizens will retain ownership, control and privacy of their data. In the same way, the engine will work with sensitive model data that the officials could not disclose for reasons of security or privacy (or commercial interest in the case of a private company).

As the world moves towards greater automation of various aspects of modern life we see the possibility of enhancing the decision-making process with artificial intelligence to serve the common benefit. This is our motivation for starting the Consensus AI project — to improve our collective coexistence by helping all of us make the choices that will lead to positive outcomes for the greater good of all. We hope you join us.

To find out more, please follow us on Twitter and Telegram and sign up to our mailing list on Consensus AI website.