Technology isn’t Neutral. Even Decentralized Cryptocurrencies.
The cryptocurrency community has been in a tizzy over tax reporting language in the Biden Administration’s infrastructure bill. There are important matters at stake. As I’ve pointed out, though, the ultimate issue isn’t what most of advocates are focusing on. It’s the regulatory challenge of Decentralized Finance (DeFi).
(For those unfamiliar with DeFi, the Wharton Blockchain and Digital Asset Project and the World Economic Forum recently published a summary report. For those too lazy to read 20 pages, here’s a short, accessible overview I wrote for The Conversation. If you’re too lazy even for that, it’s about delivering financial services — trading, lending, asset management, etc. — as decentralized applications on blockchain networks.)
DeFi services are decentralized, non-custodial, and fully implemented in open-source software. That’s what makes them so powerful, so innovative, and also potentially so dangerous. Already billions of dollars are sloshing around in the DeFi world, with the prospect of much more in the years to come. Finance is already highly digitized and automated; capital flows fast when it sees an opportunity. That’s exciting — DeFi promises major benefits, including for financial inclusion as well as for professional traders — but we need to consider the risk side of the equation as well.
Both the benefits and the dangers of DeFi, like all blockchain-based systems, flow from its decentralized structure. Which is why this comment by Brian Brooks in a piece for Fortune struck me. Brian, former acting US Comptroller of the Currency under President Trump, just abruptly resigned as CEO of the US arm of Binance, one of the largest global cryptocurrency exchanges.
Brian is a smart guy, but this is a category error, and an important one. The Bitcoin network doesn’t have intent to engage in legal or ethical violations. But that’s not the point. Neither do Facebook’s nor Google’s algorithms. And you could say exactly the same thing about the (centralized, old) SWIFT network for interbank payments. It has no desire to enrich itself. And the vast majority of the time, it acts neutrally and correctly. This has nothing to do with decentralization.
In digital policy and AI ethics, it has been a truism for years that “technology is neutral” is the worst possible starting point. Algorithms and code are made and used by humans. Human imprints on source data influence them even when they aren’t.
That a DeFi smart contract can’t engage in insider trading or market manipulation, because it mechanically and transparency executes pre-defined rules, isn’t the point. The question is whether humans can use that smart contract to do so. (Narrator: They can.)
Decentralization may even make things worse. Amazon found that a hiring algorithm systematically discriminated against women. Not because the coders or code were sexist; it picked up historical biases in the data. So Amazon tried to fix it. When that failed, Amazon shut it down.
Making the system “unstoppable” wouldn’t fix the problem. It would PREVENT the fix.
This doesn’t mean that DeFi is identical to traditional finance (TradFi). It certainly doesn’t mean DeFi should always be regulated like TradFi. (As an aside, lumping together a $20 trillion sector ranging from humble credit unions to exotic global fintechs as “TradFi” is not always enlightening.) Math is hard; so is devising effective regulatory responses to innovative technologies.
The important aspect of a “properly decentralized” system is that it doesn’t matter whether the humans have bad intent. As a miner, I can set out to double-spend to cheat the system and…I will fail. (At least, if it’s a successful consensus system like the Bitcoin network, which is hard to pull off.) Market participants and regulators don’t need to ask whether I’m honest, because it’s immaterial.
Creating decentralized systems with effective mechanism design is among the key achievements of crypto. That doesn’t mean it’s a solved problem, or that every crypto system is perfectly trustworthy. My book, The Blockchain and the New Architecture of Trust, is an extended analysis of trust and how it is manifested — or undermined — in the blockchain and cryptocurrency context.
There are all kinds of reasons why blockchain systems may not be as decentralized as they seem, or have exploitable flaws and bugs, or simply push the mechanism design challenge elsewhere (as we’re seeing now with MEV game in Ethereum mining). Crypto regulation “growing up” means descending from platitudes to tackle the specifics. That’s one reason making important decisions via rushed amendments to largely-unrelated legislation is sub-optimal. But you knew that already.
All of this shows how crypto, and in particular the legal, regulatory, and governance debates over crypto, aren’t bizarre sideshows from the great issues of the day. They are intimately tied up in the foundational questions about the world that is developing. For a deeper dive on those questions, my latest book, After the Digital Tornado, is an edited collection of brilliant scholars on how to make a world of networks and algorithms consistent with human flourishing. (Available via free open access!)
I hope some of those who have been fighting the naïve “tech solves the problem” notion (when often tech creates or worsens the problem) for years in the tech policy and AI domains can help us out in crypto. We could use it.