Resume Reading — The Robot Economy Will Run on Blockchain

Close
 

The Robot Economy Will Run on Blockchain

What finance will look like when it is controlled by machines.

Our future will be bright, fast—and full of robots. It’ll be more Asimov than Terminator: servant robots, more or less similar…By Aleksandr Kapitonov & Ivan Berman

Our future will be bright, fast—and full of robots. It’ll be more Asimov than Terminator: servant robots, more or less similar to us. Some will be upright androids, but most will be boxes filled with computer chips running software agents. And there will be a lot of them. Forecasts predict that, within just three years, we’ll have 1.7 million robots in industry, 32 million in our households, and 400,000 in professional offices.1

Robots will begin to run our factories. Autonomous sensors will monitor infrastructure. Robots will order parts for themselves and raw materials for production. Logistics will be run by chains of unmanned vehicles stationed at autonomous bases. Factories will communicate with each other. Drone traffic control systems will request weather information from meteorological stations belonging to other companies.

All of this will be based on the exchange of information. Not just technical information—robots will need to develop and maintain economic relationships. Whether for a parts order or a service agreement with another company, many aspects of their work will revolve around currency transactions. Human operators will be too slow to oversee these transactions, which we can expect to happen at 20,000 transactions per second (assuming there is at least one robotic device per person). Therefore, for the future we are building, we will need to invent not just robots—but robot money and robot markets.

Isaac AsimovShutterstock


Like with any other economy, the robot economy (or robonomics) will need to solve the problem of trust. It might seem that the very act of carrying out transactions in the digital world is a solution to the trust problem. Unfortunately, that’s not the case. Automation can help find and fight fraud, but it can also create super-efficient scam agents. What’s more, transaction costs can spiral out of control when high-frequency algorithms begin to act opportunistically. The cost of verifying that a contract has been properly executed is another problem. In the world of people, the outcome of a transaction is confirmed by the contract signatories. How autonomous agents will do that is not so clear.

A naïve solution is to create a centralized digital “bank,” just as we’ve done in the world of human transactions. For each robotic service, a centralized program would be established that would be responsible for the collection and processing of commercial information, the conclusion of contracts, the execution of transactions and the control of autonomous agents.

Capital will become the dominant means of controlling robot behavior.

The problem with this approach is that it doesn’t scale. As the number of transactions grows, so does the load on the centralized bank. This translates into higher bandwidth and computing costs, which eventually become prohibitive. In addition, a centralized network will attract the attention of scammers and hackers, and is more vulnerable to malfunctions. These problems could be partially addressed by transferring some power from the central body to intermediary bodies and building a management hierarchy. But this would increase transaction costs without providing a complete market solution.

Fortunately, there is a technology that can potentially solve the economic and technological difficulties of robot markets. It’s called blockchain.


Briefly put, the blockchain is a public ledger whose information is stored in consecutive “blocks” of information, and is protected by a consensus algorithm. Blocks can be changed only when a majority of the network agrees that they should be—or, in other words, that a change transaction is valid. Incorrect changes to the blockchain, whether from mistakes or malicious intervention, are protected against.

Blockchain was first successfully implemented for crypto-currencies like bitcoin, creating mathematically protected trade operations independent of external administrators like banks or state bodies. Then, in 2015, the Ethereum platform was launched, allowing for smart contracts to be placed on the blockchain. These are contracts of arbitrary complexity that can be verified by a public network in the same way that cryptocurrency transactions are verified. They unite into one digital object the terms of a contract, and its execution.

In our opinion, the robot economy should be built on these smart contracts. They naturally solve the issue of monitoring the fulfillment of obligations. They reduce friction among contracting parties. Information on transactions is verifiable and unchangeable. The unambiguous recording of information allows reliable reputation scores to be created. The blockchain can be organized so that network participants do not benefit from discrediting it.

When Past Disasters Are Prologue

Investment companies are fond of warning that past performance is no guarantee of future returns. Don’t rely on the past, they say, to make decisions about what may lie ahead. And when it comes to planning for low-probability events, this...READ MORE

Blockchain offers another important advantage: It can help organize how robots do their work in the first place. Experts in the field of robotics have long been exploring the problem of finding the best way for a set of robots to accomplish a common task.1 One of the potential solutions is a market mechanism, leveraging game theory, decision theory, and economic mechanisms to assign work.2,3 Blockchain can help build this mechanism, and enable the precise planning of tasks, evaluation of results, and distribution of resources.

For all its promise, a digital robot economy will face many hurdles. One is vulnerability to an attack. A blockchain network requires the existence of miners: These are nodes that generate blocks, place information into the network, and confirm transactions. In the so-called “51 percent attack,” malicious miners whose processing power exceeds that of the rest of the network can seize control of the blockchain. Other attacks (Sibyl, Eclipse, and so on) can produce a similar result. These attacks can have serious consequences for a network that controls a large amount of money—but they will be even more serious when the network controls the behavior of hundreds of thousands of robots and autonomous agents. Blockchain developers are trying to come up with ways to solve this problem, including new methods of mining and verification.

First responder In this Arialab experiment, a drone responds to a security alarm at a landfill site within minutes.Airalab

Another hurdle is the bandwidth and scalability of the blockchain. In its current form, the blockchain is very slow. While a traditional processing network like Visa can perform 24,000 transactions per second, Bitcoin can only do seven, and Ethereum 20. New cryptocurrency projects like Waves and Ripple promise 1,000 and 1,500 transactions per second, respectively. Developing technologies like Ethereum sharding or Proof-of-Authority also take aim at high transaction speeds. All of these new approaches try to strike a balance between the consensus mechanism working quickly, and the security of the network.

The scalability of data size is also a challenge. Each stand-alone autonomous agent can’t store the entire blockchain locally (at least not with current technology). It’s also not rational to place all the information related to the execution of a given task on the public blockchain. So where should the data live? Here, too, blockchain developers are working on solutions. They include creating “light” blockchain clients and distributed file systems (one example is the InterPlanetary File System).

If these problems are solved, we can begin to realize the full potential of a robot market.


Imagine a customer ordering a good through a smart contract online. Immediately, the request is sent to the internal network of a factory, and production lines begin to publish offers. Robotic agents representing different lines compete on material availability, running time, and historical performance metrics. After manufacturing is complete, warehousing, delivery, and logistics agents compete for the next stage of the order. Finally, after the order has been completed, all of the participating autonomous agents accumulate data on the work performed, analyze their own performance, and make forecasts about the future state of the market. The raw material procurement agent might decide that the demand for its products has increased, and purchase more raw materials. The logistics agent might learn that delivery through drones is more profitable than through land transport.

The advantages of robotization will be so great that we expect the majority of production, and a large portion of service work, to eventually be performed by robotic agents. As a result, the economy of robots will account for most of the total economy. Together with the rise of robots will be the advent of what we call “supercapitalists”—investors in the economy of robots, leveraging the efficiencies and scale of the robot markets. Capital will become the dominant means of controlling robot behavior.

The volume of goods and services produced by man will undoubtedly fall significantly, but not to zero. At the same time, the value of man’s economic output will drastically increase. Hand-made goods will gain luxury status, meriting a special label—“Made by Human.” Eventually, this will apply to creative activity, too. Work that involves one man or woman supporting another, and which cannot be simply automated, will start to receive governmental support, along the lines of a universal basic income. In the end, being an ordinary, honest, and aware citizen will become a job in itself, and one that we can all aspire to.


Aleksandr Kapitonov is an assistant professor at ITMO University and a member of Airalab.

Ivan Berman is a systems analyst at Drone Employee (an Airalab project).


References

1. Tobe, F. 22 research reports forecast sustained robotics industry growth. therobotreport.com (2017).

2. Choi, H-L., Brunet, L., & How, J.P. Consensus-based decentralized auctions for robust task allocation. IEEE Transactions on Robots 25, 912-926 (2009).

3. Trigui, S., et al. A distributed market-based algorithm for the multi-robot assignment problem. Procedia Computer Science 32, 1108-1114 (2014).

Issue 065

In Plain Sight

Explore This Issue

Join the Discussion