Big data has broken privacy. We’ve built a way to solve this problem and create business intelligence while preserving user privacy. This post dives into how we are automating private business intelligence.
Introduction to Ocean Protocol’s decentralized data marketplace
The Ocean Protocol is a decentralized data exchange protocol that aims to unlock the value of data and enable secure and privacy-preserving on-chain data sharing and monetization ⛓️. It provides a framework for building data-centric applications that facilitate the exchange of data assets while ensuring data privacy, security, and compliance.
The Ocean Protocol marketplace is the heart of the entire Ocean Protocol ecosystem. These data sets can be bought and sold along with access to computation on those data. What Uniswap did for Finance, Ocean Protocol is doing for data. Ocean goes beyond just merely trading data in the clear though and makes it possible to order computation on data you never see and get actionable results while maintaining data privacy and secrecy. That in many ways is the holy grail of data privacy and security!
The Ocean Protocol approach to data markets is already gaining substantial traction in industries through the MoveID consortium and GAIA-X where Datarella is a partner. Specifically, there are already industry focused GAIA-X compliant Marketplaces utilizing Ocean protocol already released in the wild (see Pontus-X)
Fetch.ai Agents + Ocean Protocol: A Perfect Match
Switching gears, you may have also heard of Fetch.ai. It provides a next-generation protocol enabling a digital world where Autonomous Economic Agents and Microagents can perform proactive economic activity. AEAs have some unique attributes. Fetch Agents (AEAs) can find resources on their own, they have private keys so they can transact autonomously, and, more importantly, they can decide to execute business logic autonomously based on their programmed skills and behaviors. AEAs are perfect for delegating on-chain tasks that you need to make sure actually happen technically but don’t want to have to handle manually. A great example of this is the stop loss agents that Fetch.ai has released in order to allow people to provision liquidity on DeFi protocols autonomously. Another real-life use case is the kind of behavior that we demonstrated with MOBIX at the International Automobile Show (IAA) in 2021.
In that context, we demonstrated how an Autonomous Economic Agent, running on an edge device and accessible through a specialized version of the MOBIX app, could enable privacy-preserving ad workflows. That sounds super complicated, but its utility is really simple.
Smart Recommendations without Data-Lake-Honeypots
Consumers have come to expect really smart recommendations in their apps based on their current context and previous actions. We’ve come to take this for granted everywhere. It extends from the autocomplete suggestions in your search field to Netflix recommendations to mobility recommendations embedded in scooter rental apps. There’s a shadow side of this intelligence though. Simply put, in order to offer that kind of intelligence someone is invariably building a vast data lake full of your user data to power those convenient algorithms. We all know that this kind of data accumulation has massive privacy ramifications, not to mention the security risks associated with those giant data-lake-honeypots. By connecting Fetch.ai Autonomous Economic Agents with Ocean Compute to Data workflows and the Ocean Marketplace we are laying the cornerstone for enabling that kind context aware convenience without sacrificing any user privacy.
Moreover, outside of the consumer sector in the world of IoT devices, this code makes it really convenient for any number of machines to automatically sell their data either directly on the Ocean Marketplace or via Compute to Data! AEAs have a certain amount of built-in intelligence. This enables IoT devices from cars to weather stations to decide, based on their situation, when and how to post the data to the Ocean Marketplace for sale. Additionally, it allows consuming devices to autonomously search out and purchase the data that they need. Together, this approach provides the basis for autonomous adaptive machine learning without having to own your own ample data sets.
Privacy-Preserving Prediction Engines
TLDR: By enabling Fetch Agents to post data to Ocean Marketplace Compute to Data recommendation engines can be driven without violating privacy.
Datarella has joined forces with Fetch.ai and Ocean protocol over the last few months to build an Ocean Connection Agent to take private data available to itself, store it confidentially, and then spin up a fixed-rate exchange on the Ocean Market or with compute to data enabled as shown in the previous figure. You can imagine that in the near future, this might be your current location data or data about your preferences or searches performed locally within an app (the MOBIX app for example). This data never leaves the sovereign control of the user and the user doesn’t have to worry about handling or organizing or authorizing the data because the AEA does that on behalf of the user well, autonomously! The same applies to companies running AEAs on IoT devices too. You might be asking now why this is even beneficial. It’s useful because it solves the problems associated with those vast data lakes that are driving the recommendation engines behind all sorts of “smart” services.
The Solution for Automating Private Business Intelligence
Ocean Marketplace is an open-source decentralized market that you can fork to develop your own marketplace. It comes equipped with all the features right out of the box. The “Compute to Data” part enables the person, or in this case, AEA is posting the data to stipulate that it cannot be downloaded in the clear – but rather that you’re selling the results of selected algorithms which you or your agent will run (automatically) against the private data on behalf of a purchasing entity (person or AEA). They get the results of the computation – the basis for those context-aware recommendations, but they don’t have the opportunity to abuse the data because they never actually see it despite gaining actionable business intelligence based on the contents of said data. That’s what we mean by “Automating Private Business Intelligence”.
The world we’re building is one where your data remains yours but the results of computations on your data will enable a host of new context-aware smart services that respect your privacy by design but still feel like magic. Bringing the unique properties of the Ocean Marketplace together with Fetch.ai Agents means for the everyday person that it’s increasingly going to be possible that you can have your cake and eat it too. You won’t have to choose between amazing functionality and privacy – you’ll be able to have both – and get paid for it. Your data – your profits!
If you’re a developer and want to dive into the code there are two spots you’ll want to check out. First of all, you’ll find the code open-sourced on github here. Additionally, if you’re already an AEA developer you’ll want to check out the package in the AEA registry. In the near future, we’re also planning to release the Storj AEA but for now, you can take any S3-compatible storage you choose. Let’s start automating private business intelligence generation!