Once the sole province of cryptographers and mathematics geeks, crypto-currencies and the underlying blockchain technologies on which they are based are increasingly finding their way into many aspects of our digital lives, and personal data protection questions are being raised as a result.
Opinion Piece by Paul Gillingwater, MBA, CISM, CISSP
According to some commentators, there is a fundamental conflict between the design of blockchain distributed ledgers, and the right to erasure (RTBF) found within GDPR.
At first glance, the conflict appears to be insoluble. How can you invoke your right to be forgotten, if a blockchain ledger, by its very design, imposes its immutability and verifiability properties on personal data contained within it?
Public vs. Private Blockchains
The key issue here is because of the design of blockchain ledgers.
In general, there are two main types of blockchain: public and private. Both types are distributed, decentralised, peer-to-peer ledgers that store digitally-signed transactions requiring consensus to synchronise changes across an unlimited number of copies.
Public blockchains are completely open, allowing anyone to participate in maintaining the ledger’s integrity. The best-known example of that is Bitcoin, which is public.
In order to verify a public blockchain, users may examine public nodes, of which Bitcoin has more than 7,000 and Ethereum (another crypto-currency) has more than 25,000.
A private blockchain, on the other hand, requires an invitation to join it, and uses a permissions model to grant access or allow even basic transactions. It’s simply invisible to anyone who doesn’t have the right access permissions.
Note that some public nodes may not be accessible (as they may be protected by a firewall), but that does not make them private blockchains. And full nodes cannot be private by definition, as they must accept external requests, often on a 24×7 basis.
Encryption and Verification
Most proponents of blockchain understand that the underlying technology uses encryption techniques, without necessarily realising that this does not endow them with the property of unreadability, i.e. protection against unauthorised disclosure.
That’s because the primary uses of encryption in blockchain is private keys that are used to validate or sign data, in order to provide objective proof of certain relationships, such as links between adjacent blocks. This is not the same as confidentiality. Therefore, simply using blockchain to store personal data in no way guarantees its confidentiality.
The design of blockchain means that all transactions are preserved going back to the moment that the chain was first created. You can’t edit a blockchain – all you can do is delete it (and all its thousands of copies), or append to it.
Personal Data in a Blockchain
Of course, most people who are implementing blockchain solutions that may store personal data (as defined by GDPR – see here) realise that the blockchain itself is a foolish place to store such information.
Instead, they can simply store an index and a hash. The former is a unique key that points to the relevant personal data records in a separately managed database (e.g. MongoDB or Oracle), while the latter is an objective proof that the referenced data is the one linked to from the blockchain.
Such a solution can be public, whilst still retaining confidentiality over the private data, by storing it separately, and using tokenisation. (This is the best practice with PCI/DSS, where credit card PAN data is tokenised in a highly protected system, before being usable in a system with fewer controls.)
Additionally, this potentially solves the issue of the right to erasure, as the personal data can be permanently deleted from a classical database, without causing any integrity issues whatsoever with the blockchain that references it.
Who owns a given blockchain?
This is a very interesting question, because as far as I can tell, once a new chain has been launched, it’s owned by no one apart from the individuals who are able to execute a proof-of-work and sign transactions.
There’s no single location where a blockchain could be deleted, as popular chains have tens of thousands of copies, each faithfully following approved changes.
Who does the approval?
It’s not centralized, so anyone who can convince a majority of nodes that it has done the required work and holds the respective secret keys, can add new transactions, such as transferring Bitcoin parts to other key holders.
This brings up an important security consideration: blockchain currency systems such as Bitcoin and Ethereum rely on a simple majority consensus for its trust model. That means if more than 51% of the mining capacity can be compromised by a single individual (which is unlikely for all but the smaller coin systems), then it can easily be rewritten in a double spending attack, thereby defrauding the other investors in that particular chain.
The founder of Bitcoin apparently owns 980,000 Bitcoins from his original mining, none of which appears to have been touched – making the founder worth over $6 billion at current rates, of which not a penny has been spent.
Essentially, being distributed with no apparent single point of contact means that data subjects have no one to contact about their rights. Data retention policies may also be violated, since there is no lifetime limit to data – once it’s in the chain, it’s there as long as copies remain in the wild – and if it’s public, then anyone can make an indefinite number of copies, with no control over who and why.
Amusingly, this means that the right to data portability is 100% guaranteed, as anyone can make a copy of the public blockchain any time they wish.
Privacy and Confidentiality
It’s apparent that blockchain is far from a magic bullet in respect to its ability to protect privacy and confidentiality.
That’s not really what it’s for. At best, it should be seen as a distributed ledger, with the ability to mathematically prove the integrity of a chain of transactions, or even smart contracts that depend on certain conditions. While confidentiality can be achieved (e.g. by using a private chain), again that’s not the primary intent of using this technology.
It could be argued that blockchain includes Privacy by Design and Default. According to S. Nakamoto’s original paper, “Bitcoin: A Peer-to-Peer Electronic Cash System“, privacy of transactions can be achieved by keeping public keys anonymous.
Furthermore, new key pairs should be used for each individual transaction, to avoid linking them to a common owner. Furthermore, Bitcoin’s very design precludes the use of a trusted third party because of its peer-to-peer nature, thus enforcing a higher level of privacy by eliminating third party reliance.
Crypto vs fiat
Technology considerations aside, there are other privacy-relevant questions pertaining to the trading of crypto-currencies.
In general, many currency traders fall into two major categories: pure-crypto, and fiat traders (This has nothing to do with the Italian brand).
The former means that they have elected to trade only crypto-currencies for other crypto-currencies, for example, trading Ethereum for Monero. The latter means that you are trading digital assets (such as crypto-currencies) for “real” money, i.e. fiat currencies such as pounds, dollars or euros.
In general, most trading platforms see that pure crypto trading is largely unregulated, and therefore may be carried out fully anonymously.
This usually means that the trading platform will collect the minimum info needed to permit the trader to login, such as an email address (which will usually be anonymous if they’re careful), source IP address (often linked to a VPN end-point or via TOR), and a unique identifier, such as an index or public key used for validation of signatures.
Each of these items of information may be considered as personal data – however, in the majority of cases, there will be a very low probability that the individual may be identified, assuming they are following recommended crypto trading security protocols.
Identification of the Owner of Bitcoins
One issue associated with Bitcoins, and probably most other blockchain-based instruments, is that by design, the system allows the owner of Bitcoins to remain anonymous.
For this reason, crypto-currencies are often favoured by criminal elements – for example, many ransomware tools request payments to be made in Bitcoins. Of course, exactly the same argument can be made about cash, i.e. that it’s preferred by criminals. (Privacy proponents want the same thing, but without the criminal intent.)
The key difference is that cash cannot be sent anonymously over the Internet, especially to cross international borders and evade currency controls. Bitcoins and their like can be sent via email, or written on the back of a postcard.
One key reason why criminals prefer crypto-currencies is that the beneficial owner cannot be easily traced, especially if they perform trades through TOR gateways.
Of course, a highly motivated intelligence agency with access to banking records, and the ability to send subpoenas or national security letters to payment gateways and Internet providers, can potentially penetrate the layers of security, as shown by the arrests of many international hacking gangs (where extradition treaties are effective, of course.) The NSA has developed tools such as FoxAcid and QuantumInsert to help with this task.
Related to this is an interesting question: can you prove you are the owner of a set of Bitcoins?
The answer is a very clear yes, and this would be cryptographic proof, which is largely unassailable. Simply transfer part of the coins to another individual, and QED.
It’s even possible to sign a non-transfer transaction, which has the same effect, i.e. proving that you can sign transactions using the private key that matches the public key associated with a specific Bitcoin wallet.
However, there is a problem – it is possible for multiple individuals to have access to the same private key, which means that it is impossible to be sure that a private key is associated solely with a single individual.
Money Laundering Issues
Therefore, from the perspective of the trading platform, crypto currency traders may happily trade anonymously, but will not be permitted to exchange their 1s and 0s for dollars.
So why this restriction? Because most countries have laws that regulate the transfers of fiat currency (i.e. dollars, yen, euros, pounds, etc.), and in particular are enforcing regulations to monitor, detect and prevent money laundering, especially for amounts greater than $10,000.
The two primary regulations at work here are the E.U.’s 4th Anti-Money Laundering Directive (4AMLD), and the U.S. Money Laundering Control Act of 1986 (Public Law 99-570, specifically 18 U.S.C. § 1956 and 18 U.S.C. § 1957).
Together, and with a few other laws in other jurisdictions, most trading platforms are obliged to perform enhanced due diligence on people who are trading crypto-currencies for fiat money.
This means a significant escalation in the amount of personal data they must collect, including copies of identity documents, utility bills (evidence of Proof of Address), full name, address, email, telephone number, etc. From a GDPR perspective, this is quite straightforward, and is only indirectly related to blockchain.
The fact is, due to the peer-to-peer nature of crypto trading, it’s entirely possible for two people to meet up in real life, and exchange currencies (including fiat) at any rate they may choose, and with zero supervision, just as people may do with cash.
Should You Even Bitcoin?
First, it’s important to distinguish one of the many possible applications (i.e. trading crypto-currencies such as Bitcoin) from the underlying technology (blockchain.)
My recommendation for small to medium businesses is a resounding yes, you may potentially benefit from making experimental investments in the technology. (Most large companies are already investing in this tech.)
At a minimum, prudent businesses will identify the risk of being hit by cryptographic ransomware, and there are rumours that some of them are already stockpiling Bitcoins (or other flavours) in order to limit the cost of paying ransoms.
Apart from this consideration, some scaremongers have suggested that relying on blockchain could lead to a violation of GDPR, and thus increase the risk of a regulatory enforcement action, that can be very expensive. In my view, such claims are overblown.
Yes, your basic Bitcoin or other currency may contain elements that are considered to be personal data, and the RTBF is hardly possible due to blockchain’s design, but these design deficiencies are not likely to lead to fines by themselves.
The real issue would be to what extent the technology is being abused in regards to personal data, and that’s really a question for the ecosystem around it.
Consent & Blockchain
One area where blockchain may be beneficial is in recording a publicly verifiable and provable evidence of records of consent.
Several vendors are offering tools in this problem space, where data subjects’ consent is tracked using a blockchain system, which also tracks the potential withdrawal of that same consent.
As long as the personal data of the individual is not encoded in the blockchain itself, but some pseudonymised identifier is used that can be referred to by the controller, then blockchain is a perfectly reasonable technical solution to tracking consent.
Performing a DPIA
When considering the introduction of a new method of processing that is based on technology such as blockchain, supervisory authorities strongly recommend the carrying out of a Data Protection Impact Assessment (DPIA). This analysis means reviewing the necessity, proportionality and controls, and identification of risks to individuals.
Blockchain-based solutions, if properly designed, can be an effective tool in maintaining data protection principles, as long as the storage of personal data is kept to a minimum within the blockchain itself (i.e., the public key.) In a way, the public key may be considered as pseudonymised data, where the index is held directly by the data subject.
From that perspective, it seems reasonable to say that a proportionate use of blockchain technology that takes into account privacy by design and default principles would not constitute an excessive level of risk to the rights and freedoms of individuals.
Therefore, the DPIA may properly focus on the surrounding ecosystem, and the locations outside of the blockchain where private data may be processed.
Having said that, it’s clear that the right to be forgotten and the right of rectification cannot be satisfied using blockchain technology, as that directly contradicts the immutability of the underlying data, so the practical uses may be limited if these rights are considered paramount.
Controllers & Processors
Where things become complex is in trying to determine the Controller/Processor relationships involved in blockchain.
Because it’s a large, distributed and decentralised peer-to-peer network, my best guess is that nodes within a blockchain network are joint controllers, effectively processing the same transactions on the same data. Bitcoin wallets however are not really nodes, as they don’t process the entire chain.
In summary, there’s little overlap or conflict between blockchain and GDPR when properly implemented, but there are data privacy implications in the ecosystem around the use of such technologies, and especially the inability to exercise the Right to Erasure or the Right of Rectification.
A little research will show that despite the hype and criminality (such as hundreds of bogus Initial Coin Offerings) associated with blockchain, it’s not going away, and will most certainly have a significant influence over the nature of our digital lives in the future.
If you want a more technical analysis of the challenges of meeting GDPR requirements using blockchain technologies, I recommend this blog from Oracle.
For more on the money laundering aspects of crypto-currencies, please click here
Paul Gillingwater is an Associate Partner at Chaucer Group, responsible for privacy and data protection.