Can AI Give Legal Advice?

Written by: Teo Spengler

May 19, 2020

5 minute read

Artificial intelligence (AI) is taking the business world by storm. AI systems can ingest, filter, and react to data in a variety of ways. But there are limitations when it comes to the scope of work AI can undertake given its inability to “think” in an abstract manner.

Can AI give legal advice? The current answer to that question is “no.” In the long run, however, the answer to the legal AI question becomes a very lawyerly “it all depends.”

With the popularity of automated assistants, it’s not hard to imagine a scenario in which AI could give legal advice: “Hey Alexa, is my prenup valid?” or “Say Siri, that neighbor is making late-night noise again. Can I sue?” While this could be a quick, easy and inexpensive approach, we aren’t there yet.

A number of legal websites use AI to answer some types of legal questions, but legal information and legal advice are two very different things. Technology has not yet developed to the point where AI legal advice could replace that of human attorneys.

Artificial Intelligence Differs from Human Intelligence

The capabilities of AI and human intelligence remain hotly debated. AI excels at many tasks once performed by humans, such as translation and contract review. While some people fear AI could make human workers obsolete, experts call this unlikely. The more likely reasoning is that AI technology will continue to work effectively with humans, not replace them.

When humans collaborate with AI, the intelligence forms enhance each other. Some skills humans excel at are simply not possible for machines. Take applying common sense to a set of circumstances, for instance. So far, AI cannot perform abstract thinking or transfer knowledge from one area to another.

And the things machines do best, such as rapid data analysis, are beyond the reach of humans. Human and artificial intelligence complement one another, and the law requires both capabilities.

Computers can assist human workers if humans train them to perform a task required in a particular job. But human workers are necessary to translate the AI results and explain the results of the AI-performed jobs to others.

For instance, the European Union’s General Data Protection Regulation requires that, when companies make decisions based on AI algorithms (such as which rate to offer on a credit card), human workers explain those decisions and their basis to consumers.

Explanations are particularly important in evidence-based industries like the law. An attorney must explain to clients how AI weighs inputs and reaches conclusions. While legal AI cannot apply a result to particularly complex circumstances, lawyers can.

AI and the Practice of Law

All states make it a crime for someone without a current law license to practice law. For example, paralegals who prepare divorce papers without attorney supervision have been found to be engaged in the unlawful practice of law.

Those who market AI tools to prepare legal documents have also been chastised and, in some cases, their activity has been criminalized. Yet today AI legal service providers offer a range of services which include automated document assembly, electronic discovery, legal project management, and legal process improvement and data analytics.

Some law firms even offer self-service portals where their clients can receive legal advice 24 hours a day with systems that consider legal issues through a series of branching questions and answers. Other sites feature chat boxes powered by AI which provide legal information. But issues can arise when using AI in the law, and AI can’t offer everything humans can—including the attorney-client privilege.

AI and the Attorney-Client Privilege

The attorney-client privilege protects confidential communications between a lawyer and a client made for the purpose of obtaining or providing legal advice. And the opinion-work-product privilege attaches to work that includes an attorney’s mental impressions, conclusions, opinions, or legal theories.

In some countries, such as New Zealand, the communication privilege only applies when a human legal advisor is involved, so AI “advice” would not be covered. In the U.S., the issue of whether “chat box” legal advice generated by AI falls within the attorney-client privilege remains uncertain.

Another more basic question is whether or not AI should stick to providing legal information or expand into giving legal advice. Legal information consists of hard, cold facts that can answer legal questions with a fixed answer. For instance: What is the waiting period for a divorce in California? Does a will in Florida have to be notarized? AI can tell you the answers.

AI can communicate legal rules and apply them to a person’s circumstances in very simple cases. A computer can, for example, tell you how old you have to be to vote in New York and, by asking the person to input his age, advise him whether he can legally vote.

But AI cannot determine the best course of action for an individual in a complex situation. Legal advice is just that: applying the law to an individual’s complex circumstances. Is it best for her to sue? To fight charges? To consider a settlement? To breach the contract and take her losses? For these questions, AI cannot tell you the answers.


While AI can and does improve the efficiency of attorneys, it cannot replace them. Clients develop a personal relationship with their attorneys. This trust builds over time when working with a lawyer who—with compassion and empathy—steers a client to the best possible result in legal situations. Most experts do not foresee AI developing this capacity very quickly—if ever. For now, attorneys can work with AI software to save time, improve efficiency, and leave room to work with clients on a human level.