Whether you are an avid consumer of legal technology, attempting to leverage the latest tools in your favor, or simply hoping to avoid messy eDiscovery disputes before retirement, a certain level of baseline competency around Artificial Intelligence (AI) technologies will serve you well in the coming years. That’s because whether you choose to employ algorithms or artificial intelligence to assist you in your review process or not, chances are your adversaries already are or will soon be leveraging these technologies.

I have countless real-world experiences utilizing emerging technologies in legal applications and I offer this article to help you take advantage of the power of Artificial Intelligence (AI) in your practice.

Let’s begin by defining and examining different models of artificial intelligence, how they can make your document review more efficient and well-informed, and why you can’t trust them completely.

What is Artificial Intelligence?

Artificial Intelligence has many definitions, but the most stringent involve a computer or program capable of teaching new information to itself and solving complex, multi-layered problems without human intervention. The pathways and programming code they use to arrive at correct answers are oftentimes beyond our comprehension, making them black box technologies, meaning we don’t know how or why they work, but they do.

In the legal world, we have been a bit more generous as to what we label AI. Complex and powerful algorithms don’t meet the purest definition of AI, as they don’t program themselves. That hasn’t stopped marketers from calling them AI for almost 20 years. With the imminent addition of “ChatGPT style” generative AI to most document review platforms, the line between advanced machine learning algorithms and true AI is so blurred it no longer matters.

The History of AI

Artificial Intelligence has been around since the 1950s. Claude Shannon designed and built a mechanical mouse that was eventually able to find its way out of a labyrinth, without any interventions.  The mechanical mouse could also recall the course it had taken to escape and apply that for future use. This is largely considered the first successful experiment in AI. Not long after, newspapers were reporting on the advancements of the “electric brain,” a machine that was promised to someday do your thinking for you.

In the decades since, much has changed, yet the elusive AI revolution has always felt just out of arms reach, or in Silicon Valley speak, “ten years out”. Multiple generations of investors have horror stories about investment dollars lost to AI “sure things”.

Given how much investment has flown into the sector recently, I’d suggest those days aren’t over. The avalanche of investment flowing into AI is only going to increase the quality and power of our options, while the large number of players in this industry should keep pricing competitive.

Stay tuned to learn more about eDiscovery and AI, including the beginnings of AI and document review with TAR 1.0 and today’s Continuous Active Learning (CAL) Models.