Everything Legal Teams Need To Know About AI 

It’s changing the legal profession

New artificial intelligence innovations are arriving daily, automating tasks that could previously only be performed by trained human beings – and legal professions frequently top the list of jobs that AI will soon automate away. While a fair portion of this is media sensationalism, artificial intelligence is nonetheless poised to have widespread and significant effects on the practice of law, which means lawyers and legal professionals need to have a basic understanding of AI. In this eBook, we answer the three most common AI questions that lawyers and legal professionals ask, so you can get a step ahead of the artificial intelligence curve.

What is AI?

If you pose this question to a computer scientist, a marketer, and a philosopher, you'll get at least three different answers. AI, ‘artificial intelligence,’ is the umbrella term for a range of techniques that allow computers to perform complicated tasks before only performed by humans. There are many different types and subsets of AI, however, machine learning is the most common. Machine learning works by taking an input, learning from it, and producing an output. These models improve over time as they’re exposed to more and more data. 

As with all automation, AI takes a lot of upfront work to build the mechanism that can perform the automated task but, once built, the system can make jobs that were formerly manual and tedious very efficient. 

With artificial intelligence, tasks deemed impractical or impossible to program are now realities. One of the clearest examples of this is computer image recognition, which powers products like Facebook's photo-tagging, Google image search, and the navigation systems of self-driving cars.

Manually programming a software system that can determine whether a photograph depicts a person, let alone identify the specific person in a picture, is virtually impossible. People come in all shapes, sizes, and colors; wear an incredible variety of clothing; and are shown from a dizzying array of angles and viewpoints. Writing a set of hand-coded rules that can account for all these variables would take decades, if not centuries – assuming it was possible at all. 

Instead, AI developers created a very basic set of programs that could break down photographs into patterns and shapes. They then fed those software algorithms a series of photographs – some depicting a variety of people, some not – so the software could write its own "rules" for distinguishing between pictures with human beings in them and pictures without them.

By feeding billions of images so algorithms like these could constantly improve those rules, we now have hundreds of software solutions available today that can recognize both human beings and human faces in pictures and images. If these systems have the right training data, they can even identify specific people shown in videos and still images. This previously impossible technology is now so well understood that it comes standard on any modern smartphone.

Now, imagine what algorithms like this could do when, instead of pictures, they’re fed decades of case law and sample legal agreements, as well as all the academic analysis that separates good contracts from bad ones and valid legal decisions from those that were overturned.

You get an AI system that can do lots of the same legal analysis as an actual lawyer. 

What it can (and can't) do

Contrary to what you see in the movies, we're a long way from disguised Terminator robots arguing cases in courtrooms. To date, we’re nowhere near developing artificial general intelligence, which is an AI that can take lessons learned from one task and apply them to another field in the same way humans do. However, there have been a ton of recent breakthroughs in AI that are quickly transforming many industries, legal included. 

An artificial intelligence algorithm that can identify human faces isn't also able to, for example, identify models of cars, identify broken bones on X-rays, or identify cat pictures on the internet. There’s no generic "being good at seeing" algorithm. By the same token, there’s no "being good at thinking" algorithm, so no AI can replace everything a competent legal professional can do. 

Artificial intelligence systems are specialists, and the kinds of things they specialize in fall into several general categories: classification, automation, prediction, and, more recently, generation.

AI classification

The image-recognition examples we've discussed above are examples of artificially intelligent classification algorithms. These systems analyze data, identify its component parts, and sort it into categories. "Does this image depict a person and, if so, who?" All other AI functions start with classification.

Properly trained AI can perform several basic legal classifications, like distinguishing between employment contracts and lease agreements, as well as identifying key components in a contract like relevant parties, governing law, execution dates, and termination clauses. And these AI systems can make those distinctions even when the contracts are poorly labeled, poorly organized, or poorly written. 

AI automation

Once an artificial intelligence solution has identified and classified a set of data, it can go on to take some automated actions based on that data. The most well-known example of this is with voice assistants like Amazon's Alexa or Apple's Siri. They can analyze the content of a voice command, then convert those words into pre-programmed actions. So, when you ask Siri to make an appointment, it can parse your statement for key values – like times, places, attendees, and meeting topics – and use that to automatically create a calendar event.

Similarly, a legal-centric AI could help speed up redlines by identifying key clauses in an agreement and then automatically flagging the sections that deviate from your boilerplate language. Advanced legal AI can even parse key values – like renewal dates – and automatically generate reports or lists for your sales and marketing teams to use to drive customer retention campaigns.

AI prediction

Artificial intelligence can classify and identify more than just text and images; it can identify patterns in data. Once an AI has identified a pattern, it can predict what’s going to happen next based on how those patterns have behaved in the past. AI prediction algorithms are used to create weather forecasts, power automated stock-trading systems, and predict election results.

In a legal setting, an AI could help analyze your contracts and, by identifying document types and their past execution dates, tell you which quarter you can expect the most new sales (as in, the most sales contracts were written), the most customer churn (the most opt-out clauses were invoked), and the newest employees hired (the most employee agreements executed). By parsing payment schedules and opt-out clauses, legal AI can help predict revenues. And by collating variations of force majeure clauses, legal AI can help predict your legal exposure to unforeseen events like the COVID-19 pandemic.

Generative AI

Generative AI is a more recent, explosive innovation, quickly impacting industries from legal to business and beyond. This branch of AI focuses on creating models capable of generating new content, such as images, music, and text, that closely resemble human-created content. While traditional AI models were primarily designed for tasks such as classification and prediction, generative AI takes a step further by creating original content that doesn’t exist in the training data. By analyzing patterns and structures within the training data, these models can generate realistic and novel outputs that exhibit human-like characteristics (except hands, for some reason).

There’s still a great deal of speculation as to the benefits and drawbacks of generative AI in the legal space. In its limited time in the general public’s hands, generative AI has been employed to streamline legal research, automate redlining and drafting, and has even shown promise for a natural language query interface that could take general questions and surface text. While all applications are still actively being explored, there are swirling questions around the regulation of generative AI as it takes the world by storm.

AI’s impact on legal and how to stay ahead of the curve 

If you don't want your role in the legal profession to be automated away by artificial intelligence, you need to follow these three guidelines to "AI-proof" your career.

Cultivate hard-to-automate skills

The first step in AI-proofing your job is to develop skills and provide services that artificial intelligence can't do. 

"Artificial intelligence" is not "artificial strategy" or "artificial creativity." AI software learns by example and is competent to automate a lot of legal grunt work, like reading and classifying documentation or composing agreements from a library of boilerplate language. What AI can't do is generate novel contract language or offer non-standard or creative legal advice. (And no, "creative" in this context does not mean illegal or unethical legal advice. If anything, there’s plenty of precedent for legal fraud on file that could be used to teach AI to commit malpractice better and faster than any human ever could.)

If you want to get ahead of the AI curve, focus on the strategic aspects of law, rather than the rote administrative functions. The days of padding billable hours with boilerplate contracts and routine document preparation are coming to an end.

Embrace AI first

If you can't beat 'em, join 'em. One of the best ways to get in front of the AI replacement of legal staff is to be one of the first legal professionals on your team to adopt AI tools. Once you see what automated contract analysis or AI-assisted redlining can do, you'll know what it can't do and can focus your skill development on the areas of law and legal work that AI can't offer. 

Meanwhile, your productivity will get a boost from artificial intelligence. And when AI starts forcing a reduction in headcount, you'll be one of the high-productivity legal staffers the company wants to keep.

Become an AI thought leader

Perhaps the most irreplaceable skill for surviving the legal AI revolution is an expertise in legal-centric artificial intelligence tools. If every practice is going to need AI to compete in the near future, then becoming a legal professional who understands what legal AI tools can do and how to use them is the best way to AI-proof your career path. 

Don't just use AI, advocate for AI in your practice or on your legal team. That way, when artificial intelligence comes to your workplace, you'll be one of those leaders making decisions on how AI is used and whose jobs it affects (or eliminates).

Artificial intelligence is here to stay, so get ready

Artificial intelligence is already in wide use in law practices and in-house legal teams across the world – and that adoption is only going to accelerate in the coming months and years. Ignoring AI won't make it go away, but pretending AI isn't coming will cause you to miss out on major progress. 

By understanding what artificial intelligence is, what it can and can't do, and how AI is going to impact the legal profession, you can prepare yourself – and adapt your career – to the technology that’s going to remake the practice of law both now and in the near future.

LinkSquares can help

LinkSquares offers the most advanced AI-powered legal technology on the market. LinkSquares gives legal professionals the tools and data they need to work efficiently and effectively with leaders across the company to move the business forward, faster. With solutions for contract management, eSignature, and legal intake and requests, LinkSquares is the central place for all your in-house legal needs.