the ai-lephant in the room

In the legal world there is currently a mixture of desire for AI and doubt about what it actually is within law, and whether it provides value or pure risk. I want to address some of the conjecture about AI based on experience, and how it was used at appropriate times for appropriate purposes in the year-long build process of The Brief platform.

When such a formalised industry as law has been understood and practiced within an agreed construct for hundreds of years, there will inevitably doubt over such a powerful technology entering the fraternity, and fear of that new member entering the room. Law is a status industry garnished with ego in parts and so when the newbie arrives able to perform the role of a human at 100 times the speed, for the the very attractive price of zero, there's a reasonable threat response. This is not a debate over whether AI can become an adjudicator, but what role it could, should, or can be in a practice environment.

Is it ok?

For some the admittance of the usage of AI may risk a public or private perception of laziness or cheating, but in actual fact it is a major breakthrough not just in technology but for the capability of law in a space where its permanency is undeniable.

Recently news headlines of AI usage in a legal process have been of embarrassment in both the UK and locally in Australia where high level practitioners have been caught out whilst appearing in front a judge only to find their AI usage has resulted in hallucinations that have made them look unprofessional. The tech didn't create the embarrassment, the comprehension and usage of it is what created a red face.

AI is not a problem, or the problem, you may just not yet know how to work it properly. Just like learning to drive, you start from scratch with no idea of all the mechanics that are designed for you to be in control of to successfully and safely reach an end point. AI is a magnificent tool that can genuinely help legal environments especially in a preparation phase, but only if you know what you are trying to achieve, and how that is possible given the capability of AI.

In creating The Brief it was a year long development phase that had to work in a strict linear process to where each phase had to be performed in an exact way with respect to the previous, and AI was a valuable ally in helping to build construct and methodology.

when is it Ok?

AI is not lazy, and AI is not cheating, if you know how to build a relationship with it. If you expect AI to be a silver bullet by pushing Enter, you'll be let down emotionally and professionally. AI is somewhere between a dog and a horse. A dog can be and is willing to be trained, and with the effort and patience the results create relationship harmony. And, if your a bit horsey, you'll know a horse wants you to take charge.

AI is not just Google on steroids, it's both the research assistant and the CEO if you really take the time to build your AI agent into the ultimate team member. I know from direct experience when you feel let down by AI due to inappropriate results or outcomes, you have to have that uncomfortable conversation about what you really need from it's performance. But it's not a pep talk to build confidence and get it to understand it's true potential. Wringing out its maximum capability is actually your job. Remember, AI is just ones and zeros in a computer, it's not a real person.

True successful outcomes using AI can be achieved and I'm happy to share how I've worked with legal professionals after creating a strict 4 phase process to reach sound outcomes, each with there own nuances. This is not a process of AI doing your writing for you, but designing your AI agent to be able act to build something from a start point through to a finish line within case development.

Phase one

Phase one is the Search Protocol. As an entry level phase this is a form of a Google search but is written like a piece of code in the form of instruction where exact needs and expectations are stated and committed to that enables parameters and guardrails so that the bandwidth represents an exact shape to be researching within.

How this phase one search protocol is written sets the scene for everything that follows so that the ingredients you're presented with are the foundation for an eventual outcome.

Phase two

This next phase is where the connection to legal accuracy truly begins, and that is the Sourcing Protocol.

The Sourcing Protocol is not simply a request, it's something that must be designed, instructed, and built for purpose so that your AI agent not only understands what the rules are, but it's starting to understand how the two of you will work together through the same lens.

The key to the Sourcing Protocol is making sure AI knows who, how, and what to trust based on knowledge of how you are trying to build a brief or case baseline. This will connect you to the writings and words of those who you may be appearing in front of, and at what level of proceedings.

Phase three

Phase three is the Fact Bound Protocol. This is where the due diligence from the previous two phases can use all the acquired information to this point to be used with powerful accuracy to distinctly shape an eventual case and point. The Fact Bound Protocol is the examiner marking the papers before delivering a final grade. If the Fact Bound Protocol is written with strict authority, it's the process that checks, double checks, and triple checks before offering recommendations, alterations, and suggestions to convert something that 'broadly applies', to then become something that is accurate via precedence. This is where you put your AI agent on the witness stand and conduct a ruthless cross examination.

Phase Four

Finally is the Editorial Standard Protocol. This is your identical twin. Yes, AI learns as it goes, but with this phase it learns how you have specifically trained it to learn. The result is it knows how you practice, what you are trying to achieve and will write with a tone and capability just as you would, and have throughout your career. Depending on your Editorial Standard Protocol it will produce your points within a narrative that flows like any great storytelling involving a well crafted beginning, middle and end.

AI can be your best friend or your most unreliable team mate. If you think of AI in law practice terms, you don't take on a new member of staff and lead them from reception to their desk and then never tell them what their role is and how it must be performed within a focused methodology. You must induct them, mentor them, monitor their progress, add to their capabilities and responsibilities. In simple terms, make them feel included through comprehension.

Please feel free to message me or call me directly if you wish to further discuss AI protocols, I'm more than happy to share my experiences that may help.

Authored by Campbell Dore

Publisher

The Brief

+61405458952

campbell@thebrieflaw.com.au