We need to talk about autonomous vehicles and the blockchain

The momentum behind the autonomous vehicle movement has been growing rapidly as the major technology vendors, and new players, seek to define the future of road transport.  RethinkX has estimated that by 2030, within 10 years of regulatory approval, 95% of all US passenger miles will be served by autonomous transport-as-a-service providers.  

Jeff Klei, president Continental AG North America, has been quoted calculating that as many as 54 million autonomous vehicles will be on the road by 2035.  As the number of vehicles increases so too will the volume of data.   Professor Pete Lockhart, commercial director, Roke says that by 2020 there will be 8.6 million connected features in cars and Rob Toews, founder of the Stanford and Harvard Forum on the Future of Transportation, also estimates there are up to 100 electronic control units in today’s cars.  

That equates to 100 million lines of code.  Basically, a “candy store” for hackers to troll; something the CEO of General Motors acknowledged when she said, “A cyber incident is a problem for every automaker in the world.”

Given the volume of potential data and the critical nature of automated transport it will be essential to get cybersecurity policies, tools and responses right. It should be no surprise that many in the automaker community believe the blockchain is the answer.  Since the Toyota Research Institute (TRI) announced its investment in the blockchain there have been a number of blogs and articles proclaiming its potential. Take this piece from Bruce Pon at BigChainDB.  He suggests five use cases for the blockchain:

1. Track and verify automotive parts
2. Track and verify vehicle provenance
3. Streamline the supply chain and business processes
4. Accelerate autonomous vehicle development
5. Build a mobility platform for electric vehicles

All of these proposed applications share one common thread.  They are very focused around processing transactions.  It is clear there is significant value in the blockchain being used to verify transactions, avoid fraud, confirm ownership and simplify purchasing processes.

What we should not do is confuse the fact that the blockchain is immutable with providing enhanced security and protection.  Indeed, the architect of the Ethereum blockchain Vitalik Buterin has said, “A common refrain is the idea of using blockchains to build systems where “users are in control of their own data”...In these cases, it is once again important to note that blockchains do not solve privacy issues, and are an authenticity solution only. Hence, putting medical records in plain text onto a blockchain is a very bad idea.”

Applying the blockchain to autonomous vehicles could have some value.  It could act as a ledger to validate transactions between a vehicle and a service provider; for example, an Amazon delivery service dropping a parcel at a vehicle.  It is excellent for smart contracts.  It is not ideally suited for protecting data. Some commentators say the blockchain can validate and protect data, because once it is on the blockchain it is immutable; if someone changes it we all know about it.  However, there is no way to know what parts of the data have been altered nor retrieve the originals via the blockchain.

 This is a real flaw in the argument for the blockchain in relation to the cybersecurity threat.  In practical terms the blockchain does not store data on the ledger, rather it uses a hash and stores the actual information somewhere else.  That “somewhere else” is just as vulnerable as existing technology.  Hackers, if they can break into a database, can delete key data - or more likely hold it to ransom.  

Imagine driving along the freeway at 70mph in your autonomous vehicle to be told by a hacker that access to your brakes will be denied unless you pay a ransom?  (If you don’t think hackers can do this just check out what they did to Andy Greenberg, from Wired in 2015)

The limitations of the blockchain are also not confined to security.  There is an argument about whether it is the most efficient model.  Healthcare in America wrote that storing a gigabyte of data on the Ethereum blockchain would cost over $2 million.  If you think back to the predictions about the number of autonomous cars and the number of sensors the sheer scale of data could make these vehicles prohibitively expensive.

With the auto industry already grappling with the huge costs of artificial intelligence, especially when it comes to training systems, the blockchain could be more likely to slow down mass adoption than help.

The expense of storing data on blockchains is driven by the miners who require a fee for confirming transactions in what is an energy intensive process.  At MaidSafe we believe a better approach is proof of resource where the system simply qualifies whether the resource being offered has the capacity to contribute to the network.  This approach does not require additional data to be generated.

Above all, though, we believe the only answer for protecting autonomous vehicles is an autonomous data network.  The decentralized web allows us to develop platforms where there is no centralized point for attack.  You can then add layers of security on top, including strong end-to-end encryption and breaking up the data, so that it is not readable as individual files.  

Ultimately, though, automating the network removes human interference, the biggest cyber threat to any self-driving vehicle.  Yes, humans will set up the network, define the rules that govern it, but the network will operate without human interference.  This may sound like Big Brother on steroids, but it is the only sensible solution. If we remove humans from day-to-day decision-making, we will ensure the network adheres to a specific policy.  

There will be new threats that appear and adjustments to be made by humans, but a fully autonomous data network is the only way autonomous vehicles will have a foundation for secure next generation transportation.