I used to think AI would never replace lawyers. Turns out, I was wrong. But also: I think I was right.

We’ve all watched as artificial intelligence has evolved from a curiosity into something that’s fundamentally reshaping how legal work gets done. Here’s what I’ve realized: AI isn’t going to replace lawyers. But the lawyers who understand how to work with AI are going to replace the lawyers who don’t.

That’s not the narrative you’ll hear in most discussions about AI technology and the law. The conversation usually splits into two camps: AI evangelists suggest that technology will make quality legal services accessible to everyone by dramatically reducing costs, while traditionalists insist that nothing can replace human judgment. 

Both sides are missing the point.

AI’s impact on the legal profession won’t be a case of simple replacement. It will be a question of evolution. The lawyers who will thrive will be the ones who determine how to use these tools in a strategic partnership while preserving what makes expert legal representation genuinely valuable.

Subscribe to our Newsletter

Many are calling this approach being the “lawyer in the loop.” After wrestling with these questions during this period of disruption, I’m more and more convinced it will be part of the future of effective legal practice.

What is a “Human in the Loop” and What Place Does it Have in Law?

If you’ve spent any time around AI discussions, you’ve probably encountered the term “human in the loop.” It’s borrowed from automation and machine learning circles, and it refers to systems where human intervention remains essential to monitor, validate, and course-correct machine output. The idea is that while AI can handle the heavy lifting, a human stays engaged to catch errors, provide context, and make judgment calls that algorithms simply can’t manage.

Legal work still requires more than simply processing information or following predetermined procedures. Understanding context, weighing competing interests, anticipating consequences, and making decisions that can profoundly impact people’s lives and businesses will remain fundamental to what we do. When I’m reviewing a patent application, I’m not just checking boxes. I’m thinking about competitive advantages, enforcement challenges, and how technological shifts might affect long-term value.

Consider the so-called “Oxford comma case,” where missing punctuation in a labor contract led to a multimillion-dollar settlement. The grammatical ambiguity created uncertainty about overtime pay. Suddenly, what looked like a minor issue turned out to be a major legal and financial problem.

Could AI have caught that comma? Possibly. Could it have understood the broader implications such as labor law context, the company’s exposure, or the strategic considerations around settlement versus litigation? Not necessarily. 

This is why having a skilled “lawyer in the loop” matters. This new role isn’t just quality control for AI output. The “lawyer in the loop” is the one who understands what’s at stake, who can make connections across complex systems, and who can apply decades of experience to novel situations. 

Here’s the crucial part: lawyers need to actively engage with AI as a thinking partner, not treat it as either a magic solution or a threat to be avoided.

The Messy Reality of AI in Legal Practice

We all want simple answers about AI’s future role in the practice of law. Will it democratize legal services or create new barriers? Make lawyers more efficient or obsolete? Help small firms compete or widen the advantage gap?

I don’t think anyone knows yet, and I’m increasingly convinced that anyone claiming definitive answers is either naive or selling something.

What I can tell you is what I’m observing in my own practice and in conversations with colleagues. The impact of AI on the legal profession is real, accelerating, and far more nuanced than headlines suggest.

Where AI Actually Delivers

AI systems excel at routine, high-volume tasks that traditionally consume attorney time. They’re impressive at quickly identifying standard clauses, spotting template deviations, and highlighting areas needing human attention. 

Document review has been transformed. When dealing with thousands of discovery pages, AI can sort, categorize, and prioritize in ways that would take human reviewers weeks. Legal research tools can synthesize information from multiple sources and suggest directions that might not have been considered.

Pattern recognition is particularly valuable in IP work, since AI can quickly identify trends across hundreds of patents, offering macro views that inform strategic decisions.

The cost savings follow naturally, for both the firm and, ultimately, the client. When routine tasks take less time, clients pay less for them, freeing up attorney time for higher-value strategic work. 

The Risks of AI in Legal Industry

AI tools have a disturbing tendency to generate confident-sounding but completely false information. They’ll cite nonexistent cases and present fabricated information with the same authoritative tone they use for accurate content. In a profession where accuracy isn’t just important but ethically mandated, such “AI hallucinations” present very real problems.

I also have concerns about skill atrophy among younger lawyers. If AI handles the majority of research and drafting, how will new lawyers develop the judgment that only comes from doing that work themselves? Less need for junior lawyers also means fewer opportunities for entry-level positions in the field, which brings even more complexity to the discussion. 

There is also the issue of bias. AI systems trained on historical data reflect all the biases and inequities of past legal decisions. When these tools predict outcomes or assess risks, they can perpetuate existing biases, making it harder for new modes of judgment to progress. As lawyers, we’re responsible for understanding how our tools work and what biases they might carry.

Amidst these changes, client expectations are shifting as well. AI’s speed creates pressure for legal work to move at GPU pace. But strategy takes time. Context takes time. Judgment on high-stakes matters cannot and should not be rushed.

The Professional Responsibility Crisis 

The ethical implications that weave through everything I’ve described are not abstract concerns. They’re practical problems that lawyers are now navigating daily.

Security risks should prompt caution about which AI platforms to use with client data. Many tools require information to live on third-party systems, raising fundamental questions about confidentiality that need airtight answers before adoption.

Legal accountability is murky territory. When AI gets it wrong, whoever’s using it is responsible. We must treat every AI-assisted output as if we wrote it ourselves, because from a liability standpoint, we did. This creates an interesting paradox: we’re encouraged to use AI for efficiency while remaining fully liable for its errors.

Reliable regulations are essentially nonexistent. Bar associations issue guidance, but there’s no federal standard, and state rules are thin and varied. For now, each firm is setting its own policies on AI use in the absence of clear regulation.

How a “Lawyer in the Loop” Can Bridge the Gap

I believe that the most sustainable path forward is strategic integration. We should be realistic and deliberate about how we partner with AI while preserving what makes legal representation valuable. The tension between these two factors is increasingly taught, which is why clear regulation will be needed to move forward sustainably.

How might this look on a practical level? 

When using AI tools for patent landscape analysis, for instance, the system could quickly identify relevant prior art and categorize technical approaches. But the interpretation of what that analysis means for the client’s business strategy, the experience to understand enforcement implications, and the anticipation of competitive responses is where the expertise of a “lawyer in the loop” becomes indispensable.

The key, I believe, is maintaining active engagement rather than passive oversight. I’m not suggesting rubber-stamping AI output. Instead, I’m suggesting using AI as a sophisticated research assistant while retaining responsibility for analysis, strategy, and decision-making. A lawyer in the loop would be responsible for asking better questions, not just processing more information faster.

Final Thoughts: Partnership, Not Replacement

The question is no longer whether AI will change the practice of law. It already has. The question is whether we’ll adapt to use it wisely, in service of the values that define good lawyering.

The “lawyer in the loop” approach acknowledges that while the mechanics of legal work are changing rapidly, the fundamentals remain constant. Clients still want trusted advisors who understand their business, risks, and objectives. They want lawyers who can stand behind their advice and take responsibility for outcomes.

I’ll admit: the rapid integration of AI tools often makes me uncomfortable. The pace of change is accelerating, and we’re forced to make decisions about AI adoption without always fully understanding the long-term implications. 

But my discomfort isn’t opposition. The lawyers who I’ve watched succeed throughout my career are those who could adapt to new technologies and competitive pressures while maintaining core professional values and practices.

That’s what we’re being asked to do now – on a faster timeline and with higher stakes than we’ve seen before.

The lawyers who will thrive are those who learn to use AI skillfully, while preserving the human elements that make legal representation valuable. This is the only way to make sure that expert human guidance retains its value.

The technology may be revolutionary, but the challenge is familiar. How do we serve clients better while maintaining professional integrity? AI gives us new ways to approach that challenge, but it doesn’t change the challenge itself.

And that, I think, is something we can work with.

Michael Dilworth

Any examples are solely for educational and illustrative purposes. They do not constitute legal advice and should not be construed as recommendations for specific actions. For personalized legal guidance, please consult a qualified attorney.

This article is for informational purposes, is not intended to constitute legal advice, and may be considered advertising under applicable state laws. The opinions expressed in this article are those of the author only and are not necessarily shared by Dilworth IP, its other attorneys, agents, or staff, or its clients.