AI in eDiscovery: Where are We Now?

What do you get when you pair Artificial Intelligence (“AI”) with eDiscovery? A match made in cyber paradise. Though AI is utilized all around us in our daily lives, its integration into legal investigations is still on the rise and companies specializing in eDiscovery, including us here at LightSpeed, have implemented a form of artificial intelligence throughout our investigations. While AI is becoming more popular and broadly-implemented, the term is widely overused and can mean many different things. So, what role does AI play in eDiscovery exactly, and what does it really mean?

AI in eDiscovery

In terms of using artificial intelligence, the field of eDiscovery is already making progress when it comes to putting it into practice. However, the rules and processes on how to deal with the basics of document review, while keeping up with the explosion of information, are still being written.

For the most part, AI is currently being used to speed document review, categorization, predictive coding, and to ultimately make decisions based on sophisticated algorithms. This means that reviewers and investigators like ours are able to more quickly evaluate whether documents are related based on concepts and similar content, versus just a batch of documents that contain responsive keywords.

AI capabilities, like the ones offered by LightSpeed, give our teams the ability to identify all of the parties involved by analyzing their identities from the full-text of each file instead of only looking at metadata fields or email recipients. In essence, we can identify the involvement of anyone referenced in communication between any group of individuals.

LightSpeed also deploys eDiscovery solutions that leverage computational linguistics and natural language processing, which organize data into understandable chunks. These chunks of data, otherwise known as “entities,” inform reviewers of names and organizations that exist within their data and gives them a quicker ability to identify the who, what, where, and when. From there, our investigation teams can comb social media, news, and other sources of public data to get a better sense of behavioral patterns or involvement amongst these individuals.

Despite these strides, we’re still several years out from a fully automated AI process in eDiscovery where it will read from sources and proactively interact with digital forensic investigators and reviewers. Why? Foreign languages continue to be problematic for AI systems to comprehend and finding instances of collusion where culprits are using code names for transactions and individuals can be problematic as well. This type of complexity makes it arduous for investigators and committees to solely rely on AI for investigations. The technology will continue to advance; so in the meantime, it’s not an end-all-be-all but more of an augmented resource to aid in the process.

Where are We Now?

In a world where AI seems to be seeping into everything, the term has taken on a whole new meaning and eDiscovery companies are finding ways to explain how the process is integrated into their business model in a way that shows its many uses. Currently, the term is generally used to explain document review of records, but a much broader definition that covers enhancing the overall process of the EDRM is where we see the industry headed. As an example, using AI to tell when a reviewer is not at their optimum performance state, or using AI to tell when a forensic collection is missing an important data repository based on interviews or other indicators.

Overall, organizations leveraging AI and other forms of sophisticated technologies should ask their solution partners:

  • What are the project goals and expectations?
  • How will this technology supplement my existing workflow and what are the risks?
  • What processes are being used to validate outputs?
  • How experienced are the teams actually conducting and managing AI workflows?

AI to LightSpeed is giving our client’s power to ask better questions, become better informed, and do more with less time and cost. This means that we strive to seamlessly combine AI with our other services whenever possible. As new types of AI are developed, we here at Lightspeed continuously search for ways to integrate these tools into our workflows and technologies for the benefit of our clients.

Recommended Posts