Jargonbusting in legal tech

When I first stepped into the world of legal technology, I was confused. I thought I knew a thing or two about technology, but there was a whole category of words, phrases and acronyms that I had not heard before.

There was another category of things that I thought I understood, but their meanings were unclear, with people adopting conflicting definitions of them. I decided after a while to ask every time somebody used a word or phrase I didn’t understand: “what does that word you just used actually mean?”.

This led to some surprising results. For example, in one meeting, a sales representative from legal tech company was using the term “API” quite a lot. I asked them what it meant and what it stood for, and it turned out they did not actually know. In another meeting, a member of a technology team was talking about blockchain. I asked them what blockchain was. After a response that threw up only more jargon, I decided on a more basic question, which was “why are its use cases and why should I care about it”. The answer was even less convincing than the first.

Reasons to understand jargon

For those dabbling in the legal technology world, it is important to understand the jargon that is being thrown around. This is for two reasons.

First, it means you can follow conversations. You can’t spend an hour in a meeting talking about APIs unless you know what an API is. This is a basic matter of upskilling — and something most lawyers dabbling in legal technology will need to do, unless they work in a niche technology practice or have direct experience of product development.

Second, it means you can focus conversations properly. If somebody is using a term you know to be vague and full of fluff, it is often an indicator that the conversation needs resetting to something more concrete. Asking what people mean when they use jargon can often draw out crossed-wires and settle misunderstandings between groups at an early stage.

I only have time in this article to address four pieces of jargon I hear a lot, and even then, only at a superficial level of detail. However, this should be more than enough to help you follow and focus conversations effectively if you are starting from zero. (Apologies in advance to experts in these areas — I’m sure there are nuances I have missed, but I am going in at a basic level here…).

Artificial intelligence (AI)

Historically, a computer produces an output based on predetermined rules (e.g. if [x] is greater than [y], do this, otherwise do that). AI lets a computer determine what the output should be based on examples that are fed to it. If you feed an AI algorithm 10,000 pictures of a cat (and tell it those are cats), and another 10,000 pictures of a dog (and tell it those are dogs), this might start to “train” the algorithm to be able to determine for itself what is a cat and what is a dog. (The better phrase for AI, in this sense, is “machine learning”).

This kind of algorithm requires extremely powerful computers, which have only been available relatively recently. This is why AI is a modern-day buzzword: it has only just become possible.

If you were to play a word association game, using AI as a word, you might come up with associated words such as “robot”, “magic” or “Elon”. AI gets a lot of airtime in the legal technology industry .

Occasionally, people use AI as a mask to hide behind, in order to make things sound more impressive than they actually are. Some people choose not to adopt the definition of AI suggested above, but use it more as a synonym for “magic”. But adopting this definition means people can sell an otherwise simple tool to a technophile just because it “uses AI”.

Because so many companies phrase their objectives as “harnessing AI”, it is sometimes easy to justify buying a tool merely because it uses AI. Law firms often buy tools that reference AI in their marketing materials so that their pitches to clients sound more impressive.

I find all of this frustrating, because it often means people buy things they don’t need, at the expense of not buying things they do need — merely because one “uses AI” and the other does not.

There are two reasons why it is important to understand what AI actually means.

First, it helps you identify companies who are trying to deceive you. If somebody talks about AI, you should probe on what exact patterns are being spotted, and what training data has been used to inform that pattern-spotting. You might even find that a company does not use AI in the true sense at all, but instead uses a bunch of rules-based algorithms that produce a so-called “magic” product.

Second, it helps you understand whether the AI is going to meet your needs, how it actually works and how you might improve it. The best in this field will be able to have an open discussion with you about what training data informs the outputs — and, most importantly, what the end goal actually is. Because AI is useless unless it creates tangible value for somebody. This is one of the main dangers of AI — people get so excited about it, that they forget it is just one tool in a toolbox.

Associated buzzwords: machine learning, deep learning, unsupervised learning, neural networks, chatbots

Blockchain

Most technologies use databases to store information. These databases are owned and managed by a single company (e.g. a bank, a government agency, a technology company etc.). Blockchain is different from a database because its records are held by all participants. If a single participant tries to alter their record, they will not succeed — because there will be a conflict with the entries held by everybody else.

Blockchain has been championed as a “zero-trust” way of storing and transmitting information. Whereas a bank could theoretically alter individual ledgers of its customers, this would not be possible in a blockchain system. This is why blockchain is associated with cryptocurrencies such as bitcoin. Your holdings are stored on a tamper-proof ledger, taking out intermediaries such as banks. Also, the ledger can never be retrospectively amended. These capabilities of blockchain are referred to as “decentralisation” and “immutability” respectively.

There are many instances of technologies being powered by blockchain even though there is probably no need. In my view, the use cases for having a decentralised way of storing information, as opposed to a centralised way, exist only in very specific circumstances. Such a situation might arise where anonymity is important, and where it is important that no single actor can have the power to amend records.

My tip here: if somebody is talking to you about blockchain, ask the question: “why does this require blockchain instead of a traditional database?”. If you don’t understand the response, the chances are it isn’t a great one, and, as with AI, somebody might be trying to make something sound more impressive than it actually is.

Associated buzzwords: smart contract, decentralisation, cryptocurrency, metaverse

Cloud

Since business started using computers, technology such as email servers, office productivity software and intranet solutions have been deployed “on-premises”. Companies have server rooms, often housed in their headquarters, that store all emails and data produced and collected by the business. Law firms are no different.

Operating in this way means you have to spend significant amounts of money buying, maintaining and securing physical servers and software installed on them. For example, upgrades to software have to be done manually by a person with expertise. This might cause significant downtime for users. Servers and hardware have to be regularly replaced, as modern software has become increasingly demanding — particular those that employ complex AI capabilities.

An alternative model has emerged. Instead of companies owning and maintaining servers themselves, they buy a slice of a “server farm” owned by a company dedicated to this purpose, such as Amazon (AWS), Microsoft (Azure) or Google. This is the cloud.

The cloud means companies no longer bear the overhead of buying, maintaining and securing servers. Instead, they pay a monthly charge to a third party based on the amount of data they are using. Hardware can be upgraded easily to meet the growing demands of modern software, and updates to software can happen much more easily.

Law firms have been slow to move to cloud, for a number of reasons, including:

• Security. Some have been reluctant to delegate responsibilities they currently hold around securing data to a third party.
• Data residence. There may be concerns around where the cloud servers are hosted. Some data privacy requirements do not allow data to be moved outside of particular regions. There may be concerns around opening up clients to foreign regulators if data comes under their jurisdiction.
• Consolidation. If multiple cloud services are used, there are concerns around data being held in lots of different places, and it being hard to track down in the case of, e.g. subpoena requests or discovery requests in litigation.

My prediction is that in 5 years time, the vast majority of law firms will have started or completed their move to cloud. This is because long term, it is a cost-effective solution to hold ever-growing amounts of data, and the only way of ensuring you have the relevant computing power to harness modern software needs and service clients.

Associated buzzwords: Azure, AWS, Kubernetes

API

API stands for “application programming interface”. It is a way of connecting two programs together. This can be useful for a few reasons, usually if one application needs to take or send data to or from another application.

An API is composed of a set of endpoints. An endpoint dictates a specific piece of functionality that can be shared with another application.

There are APIs and there are APIs. Just because a piece of software claims it has an API, that does not mean it can seamlessly connect to another application in the way you want it to. For example, an application might have an endpoint to let you read documents from it, but it might not have an endpoint to let you update or save documents back.

Too often, people talk about APIs in a “checkbox” way. “Do you have an API” is a lazy question to ask somebody. Instead, you should be clear on (1) what you are trying to achieve by connecting to another application, (2) what functionality must exist to serve that use case, (3) whether the right endpoints exist to let you do this, and (4) how well-documented the API is to allow developers to build the integration.

Associated buzzwords: endpoint, REST or RESTful, HTTP Request

There is more jargon to bust in the legal technology world, but this should be a good starting point. In my experience, it is always worth questioning people when they use jargon. Not only will other people in the room likely have the same question, but it also helps smooth out differences in approach and focus conversation in there right areas.

This article was originally published in the Modern Lawyer