Accelerated eDiscovery Using Artificial Intelligence (AI)

by | Jun 9, 2021 | eDiscovery

All law firms have accepted the fact that a large portion of their discovery now includes eDiscovery. A recent study shows that over 80 percent of all cases now have eDiscovery involved. This number will continue to grow as more processes become automated in everyday life.

Sifting through all of this raw data can be very cumbersome, especially when you look at cases that involve corporations and potentially hundreds of thousands of documents to go through looking for key information.

The amount of time this can take and the associated costs of going through this electronic discovery are enormous. The only true solution is to involve an artificial intelligence program.

AI and the Investigations Process

Artificial intelligence or machine learning has become a buzzword in almost every industry. The ability for a program to start recognizing patterns, clusters, or similar meanings can make the process of sifting through thousands of documents much easier.

Lawyers already have the ability to search digital discovery documents using keywords. This type of search will produce documents that contain the word and help them reduce the search for relevant documents. But think about the potential information you may miss. This is where machine learning comes into play.

When using machine intelligence, you can use the keyword search function, and that search will be much broader.

For example, you need to undertake a search for the word bank. In a normal search, only the word bank is going to be tagged. You would have to do separate searches for similar words or phrases.

But sometimes standard search technology cannot help us. What happens when the word bank actually refers to different variations of the same word? We could have a financial institution, but we could just as easily have a riverbank, a bank of speakers, a plane banking, the bank of a board game, and a banked curve of a racetrack.

That is the reason LDM Global has built a portfolio of Accelerator AI models to understand the context of words and distinguish between multiple meanings of the same word, including understanding variations of “Bank”.

With advanced searches conducted through AI, the computer searches for patterns that are related to that keyword. The program will then include these patterns in the results. What this leads to is more information regarding the subject and less search time for the attorney.

Advanced searches can also spot patterns in words and sentences that may not at first be obvious to the researcher. When you are going through hundreds if not thousands of documents, it is hard to put together a pattern. However, advanced searches using machine learning can quickly pick up these patterns, which can provide significant assistance to your case.

Time is always of the essence in the legal field. Attorneys are always under pressure to meet deadlines for discovery, even when that discovery consists of thousands of documents. Integrating artificial intelligence into litigation support is not only wise. It is becoming a necessity.

“We are trying to find the stories in the data, including the hidden patterns and actionable insights … and shortening the time it takes to find all of the relevant documents”, stated LDM Global CEO when launching this service provider’s Accelerator AI models.

It Is A Learning Process

When you are working with machine intelligence, you need to teach the machine what it needs to look for in the records. The machine will start by retrieving relevant data from the initial keyword search. Then, based on what your litigation support partner does with that information, the program will “learn” what to look for next.

When retrieving the first batch of information, the support team will go through this information and determine what is relevant and what is secondary. Based on these selections, the machine can “learn” what you are looking for in the documents and then conduct another search.

During the secondary search, it will look for clusters of words that may have relevance to the documents that you marked as important or most relevant. It will also look for similar terms to the keyword and the documents that you selected as important. This secondary search will turn up more relevant documents to be reviewed.

Most support teams will conduct a third search once they mark the documents for relevance on the results of the second search. This may help find more relevant documents to start building the case.

Is this method 100 percent accurate? No, but it is very close. These programs quickly adapt to what is being searched for in the documents, and it helps eliminate hours or days of reading documents that have no relevance to the case.

Is Machine Learned Litigation Support Accepted By The Court Systems?

The Rules of Civil Procedure were last updated in 2015. These rules clarify how an attorney can obtain, sort, process, and dispose of electronic files obtained in a case. The rules do not, however, disqualify the use of machine intelligence when going through the discovery material.

Accurate Investigations Are Crucial To Case Success

It is a nightmare of every attorney that a critical piece of evidence is overlooked—the thought of missing that one key to a successful case can keep you awake all night. At the same time, looking through thousands of documents to find that key evidence is overwhelming.

After a while, all of the pages look the same. The information they contain becomes a blur. This is when you risk missing important information.

Face it, when you collect electronic evidence on a case, you can collect all the data from a computer, smartphone, hard drive, or server. This data is not going to only contain information about your case. It will also have people’s personal emails to their mom, to-do lists for when they get home, shopping lists, and birthday reminders.

The data will have unrelated information to the case. If it is a business, it could be all of their other client’s information, business documents, or its work schedule for employees for the week.

It is easy to understand how anyone can miss important information when they are sorting through so many irrelevant documents.

Using a program that learns from your choices about the documents you are looking for, you can cut your research time by 75 percent. Yes, that much. You can instruct the program to look for specific keywords or chains of words and eliminate even having to look at any of the other documents.

Finding Information That You Did Not Expect

Attorneys may have nightmares about missing information, but they do dream about finding the perfect evidence. Through the use of advanced search programs, this dream can come true.

The advanced algorithms that these new search programs use have the ability to “super search” digital data. They will search not only for keywords but clusters of words that are related to that keyword. They will also reveal patterns in information that are not easily distinguished when reading the documents manually.

The ability of these programs to find these clusters of information can reveal information that the attorney may not have even considered, That “needle in a haystack” is revealed.

Using Learned Intelligence Takes Practice

Many attorneys recognize the potential for machine learning when it comes to electronic discovery. They jump on board with the program and encourage their firm to utilize this tool. However, these attorneys can become disenfranchised with these programs because they have not practiced using them and see how the learning part of the program works.

Anyone who is going to incorporate advanced intelligence learning into their discovery process should experiment with the program before putting it to practical use. Start with a case that you have already finished or create a “case” from electronic documents held by your firm.

Since you already know all of the information that is contained in these documents, it will be easy for you to determine if the learning process is working with the program. Start with a keyword that you know it will find in the documents and proceed from there.

Once you have run the program and indicated which documents are relevant, rerun the program. It will search for more relevant data based on what you “taught it” by choosing what was relevant or not.

Once the program is completed, go through the documents it has selected and see how well it has chosen compared to what you actually used in the case or what you anticipated it to find. You will be amazed at how close it is to your final preferences.

Advanced Features of Programs Using Machine Learning

All law firms understand the importance of finding and redacting certain information during the discovery process. Privileged, confidential, and certain protected information such as health medication and personal identification information must be redacted before the discovery is released to the other party.

Many law firms will print these documents so that they can be physically redacted and submit the paper forms to opposing counsel instead of submitting a digital file. Now, with advanced learning techniques in your discovery program, you can digitally redact this information by using keyword searches.

Using this digital format for redaction saves time, printing expenses, mailing expenses and ensures that all of the information has been protected. The information can then be transmitted digitally with all of the other discovery.

Why Some Law Firms Avoid Advanced Machine Learning Programs

Some lawyers read about the advancements in electronic discovery processing and shake their heads in disgust. They have three main reasons that they do not want to use this type of technology.

1. It takes away the human factor of going through the case.

While the machine is doing a lot of the searching for your evidence and documents, it is not taking away the human touch of the case. The truth is, the program can find all of the documents you need for your case, but that is all it can do. It cannot use this evidence to build a case. It cannot connect the documents to any area of the case. It cannot conclude any information from the data that is produced.

All this program does is eliminate a lot of the time and strain of manually reading and searching through electronic data. You can reduce the time you spend reading someone’s email wishing their grandmother a happy birthday and focus only on the documents that help you build a strong case for your client.

2. “I just don’t get how the program works.”

We all have to learn how different programs work when we install them into our systems. This type of program is no different. This is why it is encouraged to do a test case or to with the program so that you can feel comfortable working with it on a real case.

When you get a new phone, you have to learn how the features work or download a new game to your laptop. It is all a learning process. When you see how well the program can improve your case outcome, you will be happy that you gave it a try.

3. It is going to be expensive to implement

Just like any other program that you use in a law firm, there is a cost for purchasing the system. However, the hours that you save each week by multiple employees will be more than enough to cover any costs.

These programs are designed to improve workflow, increase processing time, and reduce the number of empty hours dedicated to reviewing irrelevant documents. It may have an initial cost, but the program will help the firm save money. Many law firms actually see higher profit margins from the reduction of time in processing paperwork.

The benefits LDM Global clients experience with Accelerator are:

1. Time

We help you go from Information to Insight faster. Reading through all of the data, from emails to phone bills, to understand the story or narrative… And then map the data, plus present it in an easy-to-understand way.

2. Efficiency

We have created models, these are algorithms with knowledge of past matters that can be applied to new matters and these models can look for stories, learning all of the time.

3. Accuracy

Accelerator helps clients understand the context, ambiguity, and variability of language more accurately

LDM Global clients have access to a library of these models, but you can also turn previous cases into AI models and use them on your new cases to instantly understand priority documents, people, facts ..etc..

The Facts About “Open Discovery” Laws

At this time, 46 states now have open discovery laws. This means that the opposing party must have access to all of the information that the other party has so that a good defense can be made.

These laws were created to ensure that everyone receives a fair trial. However, these laws can place some very burdensome timelines on a law firm when they are trying to review thousands of documents gathered digitally.

There are time limits set as to when a firm must turn over its discovery to the other attorneys for review. The original firm holding this information must have all the information they need so that they are already building their case when this information is released.

When you turn over the information to the other firm, they are also providing you with the evidence and documents that they have gathered. This means that your firm must be prepared to review this new information when it arrives.

The most effective way to do this is to have a program with learned intelligence going through the documents so that you can have the relevant information from your original discovery and from the new discovery in a quick manner.

If your firm is using advanced technologies, it will also provide you with an advantage over the other firm that still may be processing discovery the old-school way. There are no requirements stating that both sides of an issue must use machine intelligence for processing the discovery documents.

In the end, lawyers and law firms must accept the fact that digital information will become the leading data that they collect on any type of case. Having a way to process this information quickly and efficiently will be necessary for any law firm to survive and prosper. The use of advanced technologies like programs with learning capabilities and digital redaction systems can significantly improve the discovery process for any law firm.

LDM Global created Accelerator where they have taken the best of available legal technology to create tailored tools to help legal teams find the most critical information fast, efficiently, and more accurately.

The best technology combined with the LDM consultant’s expertise has allowed this 25-year-old service provider to develop unique Artificial Intelligence models to provide answers to very complex problems at speed and scale.

Conor Looney is also an advisor to the Electronic Discovery Reference Model’s (EDRM) Global Advisory Council.
Connect with Conor Looney