Why Toyota thinks blockchain could enable self-driving cars
Toyota's Research Institute (TRI) captured headlines this spring when it announced plans to partner with the Massachusetts Institute of Technology and other data experts to explore how blockchain technology could speed up self-driving car research.
The gist of their pitch is that the car industry needs the power of blockchain's decentralized security features to get public and private stakeholders, who have separate profit goals, to agree to standardize regulations before automation takes off.
Toyota says the blockchain could make it easier for companies and communities to come together to analyze the huge amounts of data expected from sensors in cars, roads and other new transport devices. Eventually, this should lead to efficient smart transit everywhere. The Institute is optimistic that blockchain technology may "create transparency and trust among [car] users, reduce risk of fraud and reduction or elimination of transaction costs." Adoption also could mean profound changes for insurance companies.
It's the money, stupid
For a sense of where the blockchain could play one of its biggest roles, it makes sense to consider the technology’s origins. Blockchain's biggest success, its bitcoin exchanges, seem to be ideal neutral ground marketplaces.
Each have millions of accounts that publicly deal digital currency using private, pseudo-anonymous IDs. The bitcoin’s blockchain system, in particular, has proved financially lucrative by running a distributed peer-to-peer system through verified transactions in a public ledger and without a central regulating authority, where trust lies in the users and not institutions. The blockchain doesn’t carry fractional costs such as middle-man fees.
Fractional costs are how corporate banking systems, as well as tech companies, make tons of money, noted Chris Ballinger, head of TRI. Because banks incur most of their risk storing, protecting and verifying the value of money, they charge people what the market bears for doing so — which makes costs add up for users over time. "In a conventional bank payment mechanism, small payments are problematic because of the cost of the system," he said.
Tech companies such as Google, meanwhile, charge for "free software services" used by billions of people by gaining access to their personal information — the web's most valuable currency.
The promise of the blockchain is that it could support millions, if not billions, of smaller transactions — accepting payments for on-demand services and collecting data (as authorized) securely.
This is really the reason why companies such as Toyota want to use blockchain in self-driving cars, according to MIT professor Christian Catalini, who is currently not involved in the Toyota project. "Self-driving sharing platforms mix the cost of networking and the cost of verification," he said, in a way that incentivizes everyone to "operate on the platform."
And because individuals more easily can control data through blockchain, the thinking goes, people can trade on the value of the data and make money, not unlike owners of solar panels selling energy subsidies back to the grid.
Get your insurance
The insurance marketplace also could be transformed substantially from blockchain technology.
Insurance companies set rates today based on verification of events most likely to take place and those that already have happened. If self-driving cars eventually reduce accidents by using algorithms that are the direct result of more-reliable data, Catalini said, insurance companies will be able to "write insurance contracts on the blockchain because they can verify what happened and do it cheaper than before."
Once people are deemed less responsible for car accidents than the computers driving them, insurers and manufacturers of cars will take on more of the financial liabilities, Ballinger said. As a result, he expects insurance companies will be even more heavily involved in setting rates and will ask to see more of that data from carmakers. “Insurers have certainty about risks," Ballinger noted. "The more car data they get, the more it'll help provide feedback to developers of self-driving algorithms as a positive feedback loop."
A negative outcome of a closed platform, in turn, could be the hiding or loss of data of self-driving car accidents.
[Learn more about blockchain at VERGE 2017, from Sept. 19-21, in Santa Clara, California.]
Sharing privately, not an oxymoron
The blockchain's ability to share data pseudo-anonymously is important because companies need to analyze the complicated data coming from cars and use it to build computations that eventually will run self-driving algorithms. Toyota is partnering with other experts in data analysis beyond MIT for this reason.
The GEM blockchain system, announced as part of the TRI platform, is known for applications in health care analysis. For example, GEM's processing (or really, anonymizing) private doctors’ visits, prescriptions and drug interactions data helps medical institutions safely match risk to insurance claims and set costs of medical procedures. And it does so without exposing the personal histories of medical patients.
The system’s privacy is also important because if it is deemed by the public to safeguard privacy well enough, individuals will be more willing to use and be tracked by more sensors, the devices that’ll be picking up all that data. People may not be totally comfortable sharing data about their cars' driving habits and locations, so Catalini said this sentiment will force established car-sharing companies working on self-driving, such as Uber and Lyft, to move to open platforms where anonymous data is shared like the blockchain.
"You can imagine a more dynamic market where Uber and Lyft compete for people's rides [instead of us getting attached to their platforms]," he said.
But to hear one of the brightest cryptographic minds in the country tell it, all these promises from Toyota may just be hype.
Joshua Baron, a cryptography expert and math Ph.D. from the RAND Corporation, doesn't understand why a self-driving platform needs to use the blockchain. Toyota and other carmakers, he said, could use other cryptography-enhanced systems such as hash functions to create secure, transparent and profitable platforms. "I worry we're viewing blockchain as a panacea for all cryptographic solutions," he said.
Other ways exist to collect information securely, Baron noted. It can be achieved if "you put a hard module on a car and verify and authenticate its data." Moreover, the blockchain doesn't necessarily buy either individuals or companies perfect privacy ("Blockchain locks you into a record that [can be looked up]") nor security ("cars are made of wires, are fundamentally autonomous and can be hacked.")
"Blockchain works when all parties in a system agree with each other and there's an economic incentive," Baron said. For the foreseeable future, the thousands of actors in transportation are unlikely to agree with each other on many things, he said.