How oracles, AI and smart ledgers fit into the future of mobility
ULedger, Fetch.AI and MOBI are bringing together the data needed for smart mobility.
With most of the world's vehicle manufacturing under its umbrella, the Mobility on the Blockchain Initiative (MOBI) might be the peak industry group for blockchains in the automotive and mobility industry.
It launched at the start of May, and the scope and pace of its developments are now taking shape with a new partnership between ULedger and Fetch.AI, which is aimed at bringing together artificial intelligence from Fetch's side and blockchain systems from ULedger's side in the industry-focused real-world environment of MOBI.
The goal is to better move closer to fully autonomous roads. Driverless cars are one part of this, but an Internet of Things (IoT) backdrop on which to run the system is just as important. The idea is that devices, such as driverless cars, will be able to communicate with each other and interact with systems like streetlights, people's phones and anything else with as little inefficient human involvement as possible.
Getting the most out of AI, or new technologies like driverless cars, also means building a suitable underlying system to reduce the inefficiency of human involvement wherever possible. This is what ULedger and Fetch.AI are up to in the partnership.
Fetch.AI is providing a "smart" IoT backdrop, while ULedger is providing a way of plugging outside systems, such as driverless cars, into that IoT fabric.
- ULedger. This is a company focused on creating systems that let companies efficiently plug data from existing systems into a blockchain.
- Fetch.AI. This is a platform that aims to be the world's first "smart" ledger, tapping into machine learning systems to create what's essentially a semi-autonomous blockchain with a kind of intelligence of its own, designed to better organise other AI entities that operate on it. It's focused on creating a fully autonomous data marketplace.
"This is one of those rare, exciting chances to connect two complementary pieces of technology in a way that hugely broadens the opportunities for the users of both," says Toby Simpson, CTO of Fetch. "Fetch provides a vast digital world for ULedger's customer's data alongside unparalleled opportunities to build and use collaborative prediction models. Fetch's users and its networks benefit from a concrete solution to the data oracle problem — an interface between the Fetch world and the real world."
The data oracle problem
A smart contract is a script that can automatically and trustlessly carry out instructions as programmed.
In theory, they're perfectly predictable because they'll always behave as programmed. In practice, their behaviour depends on what kind of data they see.
For example, consider a simple escrow smart contract that's been programmed to hold funds until the year 2050. Perfectly predictable in theory, but in practice, its actual behaviour will depend on how it knows whether it's 2050.
Has it been instructed to count off 12,000 odd days, or is it checking a computer clock somewhere? Could it be tricked into dispensing the funds if someone changes the date on their computer clock to the year 2050 and feeds it that data?
Smart contracts are trustless, but they're only as good as their data. If they get the wrong data, they won't work as intended.
Oracles are designed to solve this problem by seeing the truth and then relaying it to the smart contracts – hence the name "oracle." The oracles are blockchain entities that are themselves able to operate predictably and immutably, gathering data from the specified sources and verifiably relaying it to contracts.
The problem is that, once again, oracles are only as good as the data they receive. If they get the wrong data, they'll pass on the wrong data to the smart contract.
ULedger, as a system for easily plugging data into blockchains, is intended to be an oracle for AI contracts on Fetch.
"By integrating with ULedger's 3rd generation blockchain protocol, Fetch achieves a way to verify the data from outside their global state machine and seamless integration with other blockchains, traditional database infrastructures, and open API compatibility. Any and all blockchains that need independent verifiability of off-chain data across chains can benefit from ULedger’s technology," said ULedger CEO Josh McIver.
It's not literally magic, so it can't permanently solve the oracle data problem per se, but it can come close by offering a platform that lets companies plug existing systems into a blockchain solution and create immutable and verifiable audit trails for that data.
As Simpson said, it's the interface between the Fetch world and the real world.
"Partnering with a pioneering company like Fetch is a vital step in proving how flexible and adaptive our proprietary technology really is. Fetch's concept of an open economic ledger is likely to be revolutionary in the process of moving towards autonomous marketplaces void of human error," McIver says.
Fully autonomous data marketplaces are what Fetch is all about. It aims to do this with AI entities on an AI distributed ledger, designed to buy and sell data where needed, and eventually learn how to transact data before it's needed. The gist is that these AI systems will always have a view of what users are currently requesting it to do and what it has learned about what users want. With those dual views, it can simultaneously work with explicit user demands and proactively solve problems before they arise.
With its coupling with ULedger, and its work with MOBI, the idea is to create a real-time interconnected system to make roads safer by communicating weather updates or negotiating rights of ways with other cars. It's worth noting that Ford, a MOBI member, has already patented an idea for doing the latter.
Beyond cars, Fetch's vision is to cover other transport methods as well to bring a level of intelligent interconnection to different modes of transport. Consider an AI train network that can adapt to delays and shifting demand in real time, paired with a bus system that can also automatically adjust in real time, rounded out with an autonomous taxi network that's learned exactly where to send vehicles to find the most passengers in need at any given time.
Most of the world's vehicle manufacturing power is behind MOBI and these kinds of systems, and learning AI in mobility might be the norm quicker than one might expect. Which makes sense since AI is all about doing things before people even know they need it.
Disclosure: At the time of writing, the author holds ETH, IOTA, ICX, VET, XLM, BTC and NANO.