top of page
Search
Author Contribution

Aiming AI at routine legal tasks is helpful but also raises lawyering line-in-the-sand questions

By Lance Eliot.


Some lawyers contend that they are exasperated and exceedingly demoralized by having to do routine legal work.


According to recent nationwide surveys, nearly two-thirds of professional workers reportedly indicated that they are tired of doing routine tasks and that the adverse effects include a precipitous drop in employee morale (the survey respondents included attorneys). One assumes that the decrease in morale translates into less interest in the job, poorer performance on the job, and can produce heightened turnover as workers seek other opportunities that are more inspiring.


For lawyers, the types of routine tasks notably bordering on tedium entail everyday types of court filings and other perceived legal grunt work. The consumption of “lawyering” labor for such legal services is certainly eyebrow-raising as to expending highly trained and highly-priced lawyering talent to undertake tasks that would seem viably done by less costly labor.


Of course, law firms of larger sizes are usually steeped with para-legal professionals and other such staff to try and take on the routine tasks that do not especially require attorney-level acumen. Not all law firms have that luxury.

Furthermore, sometimes the most expedient way to get something done seems to be to do so yourself. Having to convey to someone else the nuances of a routine legal task might be more time-consuming and error-prone than just shrugging your shoulders, gritting your teeth, and doing the legal chore directly.


All of this brings up a somewhat hidden assumption, namely whether we can all agree on exactly what constitutes a seemingly routine legal task. Let’s dig into that assumed notion. Imagine the somewhat straightforward act of assembling a collection of legal documents into one comprehensive set for purposes of distribution to other parties associated with a legal case.


Is this a legal task that is routine or non-routine?


On the one hand, you could claim that the task is routine if there was no semblance of legal acuity required to perform the task. If this was simply the tedious chore of collecting the needed artifacts, perhaps scanning paper-based documents for machine readability purposes, and then accumulating them into one electronic-based mega digital document, presumably the amount of legal acumen that would be needed is ostensibly minimal.


Suppose though that the choice of which artifacts to include required attorney-level legal reasoning. And that the sequencing of the documents likewise required astute legal reasoning. And so on. We seem to have veered beyond the mundane mechanical facets of this task and into a realm that does in fact entail legal savviness.


You could get a non-lawyer to put together the whole kit and kaboodle and then use lawyering prowess to cull through the collective, tossing out elements that should be removed and perhaps outrightly legally are legitimately able to be excluded. Along the way, this might also require shifting items into a different order to make the compendium more legally telling. Etc.


But would it have been more prudent to instead have started with a clean slate and undertaken the entirety of the task via the skilled hands and cognitive talents of a lawyer?

It could be that doing an arduous clean-up is going to be messier and inadvertently leave something amiss. Perhaps safest and notably sensible to use a lawyer at the get-go for this particular task.


While you ponder that pickle, consider a more evident example.


Most lawyers would readily assert that drafting a legal contract and reviewing a legal contract is decidedly a legal task. For this, certainly, you need to exercise keen legal reasoning and leverage the capacities of a lawyer.


As you likely are aware, there are online programs that nowadays perform some semblance of legal contract preparation and review, as mentioned by this legal researcher: “In the past, Americans who wanted to handle their own routine needs without a lawyer might have tried to buy a book of forms or consulted a form-processing service with limited ability to provide customized assistance. Now those customers can meet their legal needs with LegalZoom or Rocket Lawyer, frequently at a price that no attorney can afford to match” (by authors Benjamin Barton and Deborah Rhode, “Access To Justice and Routine Legal Services: New Technologies Meet Bar Regulators” in Hastings Law Journal).


If a computer can be used to undertake such a task, does this ergo imply that the task must indeed be of a legally routine caliber or perhaps even be construed as a task well below that of a routine legal nature?


Some lawyers vehemently argue that those online programs are in fact performing legal services, routine or otherwise. The nature of what is legally routine has somewhat ill-defined connotations and will increasingly be a tough nut to crack. For those that believe these online programs are crossing over into the legal lawyering realm, the odds are that a strident argument will be made that these are decidedly not legally routine efforts. This in turn opens the door to shutting down those online programs by asserting that they violate key legal precepts concerning the practice of law.


To date, these online services have clung to the insistence that they are below such a threshold: “They do not, and under bar ethics, rules and statutory prohibitions, cannot offer services that constitute the practice of law. Their disclaimers make clear that they are, as LegalZoom notes in bold on its platform, ‘not acting as your attorney,’ and ‘not a substitute for the advice of an attorney.’ Rocket Lawyer similarly declares it does not provide legal advice, but only a platform for legal information and self-help” (per the cited article by Barton and Rhode).


Laying Out The Categories of Routine Legal Work

There is yet another perspective on what constitutes legally routine work.


An attorney that day in and day out creates roughly the same legal contracts for various clients, all contracts of which are almost identical would probably lean toward stating that they are being underutilized and performing legal services in a legally routine manner (even if there were genuine legal insights required to perform the task at hand).


It might be useful to try and unpack the nature of routine legal work. In my ongoing research about Artificial Intelligence (AI) and the law, I have postulated that there are at least four types of purportedly routine legal work that vary across a spectrum:

  1. Transparent Routine – legal work involving no bona fide semblance of legal acumen

  2. Base Routine – legal work requiring minimal legal acumen

  3. Upscale Routine - requiring legal reasoning but repetitively and with little added value

  4. Non-Routine – legal work that is reasonably said to require devout legal acumen

I have covered the underlying details of these four classifications in my various writings and books on AI and the law.


The key here is that this discussion about what is or is not routine legal work dovetails substantially into the advent of AI in the law. AI is stepping up the game of using computer-based automation for performing legal tasks and raises anew the question of where we draw the line on what automation is allowed to legally undertake versus the use of a human lawyer.


You likely already know about especially popular ways of applying AI in the law such as infusing Natural Language Processing (NLP) capabilities into the arduous chore of doing e-Discovery. AI-based NLP is especially useful when dealing with large-scale narrative-oriented datasets that require extensive and otherwise laborious by-hand legal discovery efforts. That being said, the odds are that a lawyer is the one guiding and utilizing the AI. The AI is unlikely at this stage of advancement to be performing such a task on an autonomous basis.


The list of such AI-applied uses that are arming lawyers is continually growing.


Let’s dovetail this into the conundrum about routine legal tasks.


One use of AI as applied to performing legal tasks consists of lawyers using such high-tech tools. In that sense, the AI is somewhat unseen by clients and merely part and parcel of an attorney undertaking their legal tasks, perhaps akin to using a word processing package or using a spreadsheet package to get their legal work done.


Another use of AI in the law consists of making AI-based lawyering tools available to non-lawyers such as consumers (who might otherwise be construed as prospective clients for human attorneys). This is already underway: “Some see a future in which legal artificial intelligence (‘AI’) will largely replace humans in providing legal advice and drafting documents. Others doubt that AI will progress that far. But, everyone agrees that computers are already displacing human lawyers in areas like document review and assembly and will likely continue to do so” (per the cited Barton and Rhode).


You could argue that the AI usage for consumers is only targeting the “routine” legal work and therefore not a particularly unsettling encroachment on attorneys.


Indeed, tying back to the earlier points about attorneys potentially being dissatisfied with having to perform routine work, perhaps AI capabilities solves that dilemma. Clients would take on the routine work themselves via the leveraging of AI. After which, they presumably would come to human attorneys solely for non-routine work that truly invokes legal acumen (and no longer bother attorneys with routine work).


You might fervently argue to the contrary that clients ought to still bring the routine work to the human attorneys, doing so under the presumption that it is the lawyers that will use AI as part of their lawyering efforts to undertake any routine legal tasks. Thus, the attorneys use the AI and presumably can reduce their labor efforts (and costs) accordingly, parceling out the routine legal tasks to (one assumes) less

costly AI-based legal-focused tools. But we

apparently should not have clients performing legal work via AI tools alone (some would so insist).


The arming of lawyers with AI Legal Reasoning (AILR) capabilities might seem to be the preferred approach, at least as far as attorneys are generally concerned. Doing so makes their work more efficient and effective. This might in the end be a boon for clients. Clients might be able to get their legal work done more inexpensively and sooner as a result of a human attorney being augmented by AI facilities.

Not all of this is a bed of roses.


Clients are bound to have a preference at times for using AI-based legal services. Cut out the middleman, as it were. Access the AILR from any location and at any time of the day, 24 by 7. Why go to a bank teller when you can do online banking in your pajamas? The same logic could just as easily apply to conferring with human lawyers versus an online AILR system.


That would seem to be a readily apparent concern for attorneys just about everywhere.


Let’s add another twist, one that you might not have previously noodled over. One worry is that attorneys relying upon the AI to aid in doing their legally routine tasks will gradually and inexorably become deskilled at being a lawyer. The slippery slope here is that when no longer undertaking the routine legal tasks, this precariously undercuts the lawyer’s ability to perform the non-routine legal tasks (i.e., the routine and the non-routine are ultimately exercising the same “legal muscles” as it were).


Is it a necessity to do routine legal work to keep the non-routine legal skills at the ready.


Arguments pro and con are readily devised on that contention.


Here’s yet another qualm.


If AI can inexorably chew away at legally routine tasks, suppose this eventually becomes commonly viable and nothing is remaining for AI to tackle in the routine legal task milieu. It would seem blatantly obvious that the AI would have to be extended to try and perform non-routine legal work. But this presumably threatens lawyers as to displace them in a much more substantive way than does the low-level crunching of routine legal work by AI.


Yikes, even this can be turned somewhat on its head.


Some would claim that if we don’t seek to apply AI to the non-routine legal tasks, we are going to undercut what AI can potentially attain for us as a society. By somehow a priori deciding that the AI should only be used for routine legal tasks, we will limit the range and depth of benefits that advanced AI can provide. We have either wasted AI or have dogmatically shortchanged how far we are aiming to propel AI in the legal field (by unduly shortening the goalposts, one might say).


Conclusion

Let’s do a quick recap and then consider the future underlying the nature of lawyer-proffered legal services amid the frontiers of AI in the law.


Returning to the postulated levels of legally routine tasks, AI researchers and AI developers are proceeding step by step to devise AI-augmented LegalTech systems. This usually entails starting with the simplest of routine legal tasks, perhaps tasks that barely edge into any semblance of legal prowess. Next, they rachet up to the base routine level and then stretch into an upscale routine level. Finally, the trick is to embellish AI capabilities sufficiently to tackle those “creative lawyering” non-routine legal tasks.


Should you be shaking in your boots right now as a human lawyer?


Today’s attorneys can breathe a brief short sigh of relief that AI is still not far enough along to have yet risen alarm bells about AI slipping and sliding significantly into the revered legal protections of UPL (Unauthorized Practice of Law). Conventional computing has already inched into that realm. AI is going to unavoidably exacerbate that. This is a line in the sand that is going to spur a tremendous amount of legal and societal debate.

For those of you that sneer and overtly doubt that AI will advance far enough to encroach on human lawyering, I have a piece of keen non-legal advice for you. Best to not be overly smug since the march of progress in AI is ardently gunning for those non-routine legal activities. If your legal career still has legs on it, keeping aware of what AI is doing for the law would be a prudent heads-up and aid you in adjusting your legal beagle services accordingly.


Let’s put it this way, you would be wise to not get caught with your head in the sand.

 

About The Author

Dr. Lance Eliot is a Stanford Fellow at Stanford University as affiliated with the Center for Legal Informatics, a co-joining of the Stanford Law School and the Stanford Computer Science Department.


He is also Chief AI Scientist at Techbrium Inc., has been a top exec at a major Venture Capital firm, served as a global tech executive at several large firms, previously was a professor at the University of Southern California (USC), and headed an AI lab there.


His columns have amassed over 6.8+ million views including for Forbes, AI Trends, The Daily Journal, and other notable publications. His several books on AI & Law are globally recognized and highly praised.


Comments


bottom of page