Attorneys are problem solvers, often tasked with ensuring safe adoption of emerging technologies in real time. With our added duty of competence, we must always advocate for thoughtful implementation, even when there’s a temptation to save time or costs through automation. The magic problem-solving skills of legal work come through a competent human in the mix, not artificial intelligence or any other tool.
As a Gen X attorney, I’ve seen technology evolve from typewriters to tablets and can consider new AI automations with the benefit of lived experience. AI is not the “set it and forget it” solution to automating “simple” legal work. No matter how much pressure in-house counsel is under to reduce costs, automate, and prioritize speed in contracts, using AI without involving human judgment does the opposite. It increases costs due to mistakes, missed clauses, and renegotiations after someone signs onto an impossible promise tangled in “boilerplate” wording.
AI automations are the next tool attorneys will learn to use, implement, and perfect, just as we’ve done for generations moving from quill and ink, through the beloved early word processor WordPerfect, to the variety of digital tools used in practice today. AI tools are not magic, but with innovative attorneys at the helm, incorporating knowledge and experience into the process while thoughtfully crafting improvements, they are remarkably effective at scaling the creation of clear, cost-effective, and commercially aware agreements. As with any legal technology, our tools are most valuable to clients when attorneys are responsible for the substance of our creations.
Why Context Requires a Human
I remember the joyful power of using the “reveal codes” function in WordPerfect, and how it enhanced our new computer skills. Revealing the underlying format is an ideal example of what technology can do to empower and improve the quality of attorney work. Today we can use AI to perform a “reveal” function across all types of contracts, uncovering their structures and paving the way to make adjustments at scale with speed. Automated review of repeat contracts does not, of course, eliminate the need for a human attorney’s experience and knowledge.
The gains of automation don’t fall out of thin air. Smart gains require human attorneys who are familiar with the deal, the client, the counterparties, and any third-party beneficiaries, as well as the tricky ways contracts operate when they are treated as routine or unimportant. When a backdoor standstill is slipped into a simple deal nondisclosure agreement (NDA), an automation might not pick up the ambiguity-turned-land-mine. An experienced, detail-oriented human review will find the dangerous one-word addition of “negotiated” to a standard use restriction (you will only use X for a transaction regarding Y) and discuss the risk of “negotiated” slipped in before “transaction” with a business in a way that a computer can’t. It’s the attorney who turns insight into quality and action for clients, not the tool.
Checking Our Work, Avoiding the Foot Fault
As a later-in-life law student, I took my 1L writing classes in my thirties with professors who regaled us with tales of the halcyon days of learning to Shepardize by hand, sharing horror stories of classmates stealing a key reference from the library when they needed it most. By that time, though, we were learning not only how to find those tomes among the stacks but also how to use LexisNexis, Bloomberg, and WestLaw: ubiquitous tools in today’s practice of law.
Along with learning the foundations of the legal profession, we were taught to use current innovations to become more agile, not to avoid our duty to check those citations each time. Current advancements require the same duty to check our work. This May, the Florida Supreme Court issued an instruction reiterating what we know: innovations don’t replace attorney responsibility. In Florida (and I expect many states to follow), an attorney’s representation to the court is the accuracy of our work, including existing and accurately cited sources, not a disclosure or certification of what tool was used.
Because our duty of competence is not removed by our use of technology, AI automations included, attorneys must always check our work. Instead of slowing things down, AI can help boost the volume of our work, leaving time for careful citation checks in the same timeline it previously took to create the product to begin with. Another use of AI is searching past agreements and finding how often and in what form the counterparty agreed to wording your client needs. With AI, we don’t have to invent the wheel or delay the process to search vast troves of past agreements; we can utilize its enhanced search and summary capabilities to inform and improve our output.
When, Not If, Automation Falls Short
A May 2026 article in The New Yorker surveyed college professors about the push-and-pull of AI use on college campuses. One professor, Daniel Silver at the University of Toronto, discussed showing students that the C-grade “replacement-level work” that AI produces is the floor. Students must still learn how to think and create something better than the identifiably bland output of a large language model.
The same issue affects the use of AI tools in our profession. The first draft spit out by a large language model is not the quality of legal work we are expected to produce, nor the work a competent attorney would be comfortable presenting as their own, even the ever-maligned “first-year.”
In the rush to utilize AI automations, those of us in the transactional contract space have already seen such C-level replacement work come into our inboxes, sometimes in an email “signed” by an automation itself. Automations don’t capture the nuances of a commercial relationship in drafts or emails, despite the relationship often being equally important as the outcome of an early-stage NDA or engagement letter. A human in the mix who knows how to balance relationships and risk improves timing and success by reaching out with a quick phone call, confirming the counterparty’s intent, and sending back a finalized contract reflecting positions acceptable to everyone.
The benefits of AI to a law firm are not outputs to replace thought, innovation, or competence. Instead, AI is a particularly promising tool, allowing thoughtful attorneys to organize and monitor complex problems across discovery, depositions, governance, and reporting.
The Gordian Knot AI Can’t Untangle
As attorneys, our clients hire us to solve their problems, no matter the size or the complexity, and regardless of whether it’s ever been seen before. Trust like this—clients bringing us their Gordian knots to unweave—is why our profession requires competence, honesty, and a clear path of responsibility for our actions. These guiding principles have allowed attorneys to be leaders in AI adoption, within our firms and in partnership with our in-house clients. We are trusted to use our experience and thoughtfulness to implement the next big thing without dropping the proverbial balls of competence, honesty, and responsibility.
Those core values for attorneys are why humans will not be replaced by large language models: Our craft requires a human in the loop. Tools help attorneys synthesize and organize data, so we support our clients efficiently and build both the details and big picture needed to make the most important decisions. AI tools can find and summarize, but not decide, so attorneys are well placed to use those summaries and searches to expand the information available to compliance teams and provide the advice our clients count on.
With AI, our clients will benefit from information learned from the last fifty or five hundred contracts that have been through the same queue, but only if they can also depend on humans applying context and detail for today’s deal. An experienced attorney knows why a particular commercial term matters to sales, why checking citations for hallucinations and other mistakes (before filing) matters to courts, or why the deposition answers of the CFO and CIO should be checked against one another. Clients will not and should not rely on a C-minus draft or on automated review, but they can benefit from the knowledge gained from prior experience when that knowledge is funneled through attorney sign-off.
The Road Ahead
Pretending the next generation of attorneys will only learn by doing it our way is a recipe for stagnation. At the same time, our clients deserve our expert skills enhanced by AI tools, not replaced by them. Luckily for us, our profession is already constantly adapting to the ever-shifting landscape of legislation, markets moving at warp speed, and clients who rely on us to understand established law and find innovative theories to move their ideas in the world. Buoyed by our foundational duties of competence, honesty, and responsibility—attorney innovators will lead on AI, too. Even as the competence and honesty of AI tools’ outputs improve incrementally, our responsibility is the core reason why the most successful adoptions of AI include human involvement. Attorney skills and foresight identify mistakes, right the ship, and approve the final work product before our work is touched by a client or a court.
Attorneys have always used tools, from pen and paper through word processors and digital research repositories. As long as we maintain human oversight, AI is simply the next tool in our belt, not the Sword of Damocles.
