"Blockchain" proponents claim it is a game-changing technology that is applicable to wide-ranging domains, including medical records to automotive safety. Blockchain is a useful technology, but it is not the panacea it is often made out to be. It is time to tone down the hype.

I'm a research scientist at Adventium Labs, a Minneapolis-based R&D company focused on cybersecurity, systems engineering, and artificial intelligence. I looked into blockchain because I thought it might add value to our development of safe and secure medical device platforms.

In fact, a recent Forbes article claimed that health care companies should invest in blockchain. I found that blockchain is useful for some problems, but anyone who thinks it will revolutionize computer security should look again. In particular, Minnesota's health care industry should not fall for the hyperbole of the Forbes-type articles.

Blockchain is data redundancy taken to an extreme. Blockchain is a massively distributed data storage system in which every user has a copy of the entire database. Bitcoin, which burst onto the scene in 2009, was arguably the first use of blockchain to reach the general public. Since Bitcoin has no central authority and every user has a copy of its blockchain database, any user can verify the transactions of any other user. In the realm of digital currency this makes sense as a way to maintain solvency. Bitcoin prevents fraud by making it extremely expensive to fake a transaction. So expensive, in fact, that it has never happened. An estimated $50,000 per hour is spent globally verifying Bitcoin transactions.

The verification process for transactions on the Bitcoin blockchain uses a resource intensive technique called proof-of-work. For a block of transactions in the Bitcoin blockchain to be considered valid, it must include proof that a specific math problem embedded in that block was solved. That proof is added to the Bitcoin blockchain, along with the block of transactions. This ever-growing chain of transaction blocks is Bitcoin's ledger — its database of all transactions ever conducted in Bitcoin. By design, the longer a blockchain grows, the harder it is to change.

This high level of security makes blockchain attractive for use in other products. The challenge is that blockchain's strengths bring along major weaknesses. For example, performing maintenance on a blockchain is difficult. With no central authority over the data, an error cannot simply be reversed like a chargeback on a credit card. If you accidentally typed $10,000 instead of $1,000, that is too bad, the transaction is on the blockchain and it is done. In addition, substantial changes to the rules or structure of a blockchain require consensus among its users, a challenge that grows more daunting as the blockchain expands. Finally, transactions take much longer to be verified by a blockchain-based product than by traditional products; Bitcoin transaction verification takes about 10 minutes for small transactions and 60 minutes for large ones. Keep in mind, the transaction rate for Bitcoin is relatively slow. Since 2009 Bitcoin has processed about 250 million transactions. Visa processes that many transactions every two days and processes each transaction in seconds.

Before trusting blockchain-based products as a user or investor, consider the following three questions:

• Dependability — Does the product have a dependable and sustainable incentive for users to expend resources to verify transactions? I mentioned earlier that it costs about $50,000 per hour to keep Bitcoin running smoothly. Bitcoin has stayed afloat because there are financial incentives for users to spend money verifying blocks of transactions. If that financial incentive disappeared, Bitcoin commerce would grind to a halt. If an incentive is missing from a blockchain-based product, the product will not be dependable.

• Security — Does the product rely on security measures outside of its blockchain and, if so, are they at least as secure as its blockchain? Transactions on the Bitcoin blockchain are secure because they are digitally signed by the participants. The validity of Bitcoin transactions is guaranteed because the inputs and outputs are known, visible, and accessible to all parties as part of the Bitcoin blockchain. Products that rely on a mix of data stored both on and off a blockchain are only as secure as their weakest link. For example, users of the Mt. Gox Bitcoin exchange lost millions when they trusted their private information to off-blockchain entities. If the product incorporates less secure mechanisms on the way to or from its blockchain, the product will not be secure.

• Trust — Are the digital and human processes associated with the product worthy of trust? No product exists in a vacuum. Whether stated or implied, every record-keeping system from the Department of Motor Vehicles to e-mail has administrative processes and social norms that guide its use. A record is only trustworthy when its meaning is agreed upon by everyone involved; getting all parties to agree on what a contract says is not the same as getting them to agree on what a contract means. A blockchain record of the sale of a car is only trustworthy if a process exists to enforce the terms of the sale. If the product cannot account for the human processes in its operation, it is not worthy of trust.

Bitcoin got the world excited about the potential of blockchain, but it is time to slow down and look at what really makes a successful record-keeping system: dependability, security, and trust. If a project covers all of these bases, it has a good shot at success with blockchain. If it takes shortcuts or leaves topics unaddressed, take another look and decide whether you really want to rely on the product or make that investment.

T.D. Smith is a graduate of the University of Minnesota and a research scientist at Adventium Labs.