top of page
Search

Are Attorneys Afraid of Artificial Intelligence?

Writer: Ersula Rivers, JD, CPTC, PMP
Ersula Rivers, JD, CPTC, PMP
22 hours ago
9 min read

Updated: 17 hours ago

This year Halloween falls on a weekend, and with it just days away, I'm in my feelings about artificial intelligence. Yes, AI has me feeling some kind of way. Nervous. A little anxious. I wonder if I'm on the verge of nightmares. With spooky purple lights and graveyard decorations at the houses lining my street, I wonder if my mind is playing tricks on me. What’s so scary about AI?

Is AI too much? I'm generally overwhelmed by it. In my daily routine I put a computer on my wrist to track my steps, exercise time, calories, andheart rate. I look at a computer that fits in the palm of my hand to check the time, the weather, and reminders about what I have planned to do. I'm borderline addicted to that thing. I drive an electric vehicle that tracks where I'm looking, how fast I react, where I go, and so much more data.

For work I start up a larger computer and open applications that embed predictive and generative AI. I run a morning brief connected to my communications and schedule, check online engagement, and dive into serving clients. I run automations and skills; some I created and others that pull from memory on their own when the tool decides it's necessary. Most of my interactions are digital: virtual meetings, webinars, and more. Artificial intelligence is embedded in nearly everything I do, and I've had little choice about opting out.

How people feel

Some of my older relatives have shared their opinions about AI. They admit they don't fully understand it, but they believe it's "evil." Talk about a strong reaction. As much as I don't see AI as a blessing, I don't subscribe to the opposite extreme either. The teenagers in my house hate it. They have a much better grasp of what it is, and they are very good at spotting AI-generated content, like food photos on menus that have the opposite effect of being appetizing. They think it's weird and all-around not ok.

They are not alone. Pew Research Center's 2026 survey of U.S. adults, conducted in February, found Americans deeply skeptical of AI, including younger adults. More adults expect AI to have a negative rather than a positive impact on themselves and on society, and majorities think it is advancing too quickly and will put their personal information at risk. And anyone who works has heard about the threat of AI taking jobs.

The feelings are complex. Researchers who analyzed 12,000 metaphors Americans used to describe AI over the course of a year found that people increasingly describe it as warm and human-like, and that those perceptions predict how much they trust it. The dominant metaphors in that study ranged from AI as a tool to AI as a thief. Sometimes it's a helpful friend. Sometimes it's the one stealing from your wallet.

I read a post by a leading AI expert who compared how people feel about AI to how people must have felt about electricity. In that view, the feelings didn't matter much. Applying a people-first lens, I disagree. Here's why fear matters.

Why fear matters

How people feel makes them distinctly human. Feelings tie directly to thoughts and behavior. Negative feelings like fear can result in negative thoughts like rejection, misuse, and distrust. That's an overall bad experience.

Lawyers know that a bad experience costs money, time, reputation. A bad experience with AI doesn't stay with the person who experienced it. The bad experience extends to colleagues, to the judiciary, and to clients.

There's research on where the fear concentrates. A study of 10,000 people in 20 countries, summarized in Psychology Today, found that people were most afraid of AI in the role of judge or doctor, and least worried about AI as a journalist. The researchers linked those fears to a concern that AI lacks the characteristics needed to do the job effectively and ethically. Judges and doctors are roles built on judgment and trust. Sound familiar? Lawyers’ roles are built on judgement and trust, too.

I describe what lawyers fear about AI as a "parade of horribles", a long list of speculative uncertainties, with the worst things that could happen leading the way. Lawyers know the term. Let’s look at the terrible consequences  that might follow from using AI. Let’s examine the fears.

What Could Go Wrong?


Hallucinations

We've all heard the horror stories about hallucinations, the word for errors and inaccuracies that personifies AI. The technology doesn't hallucinate like it's on drugs or having a mental lapse. These tools generate statistically probable text, and sometimes probable isn't true. There may be a bug in the system or the algorithm is producing an unexpected outcome that we consider an error. No lawyer wants to produce work that contains mistakes or give advice based on a misstatement, misreading, or misinterpretation. That makes the judgment wrong and the counsel unsound. AI can be confidently, convincingly, and defensively wrong.

Law360 Pulse's 2026 AI Survey of 506 attorneys found that firm lawyers' top concern was AI giving incorrect responses. Stanford researchers who tested AI research tools from LexisNexis and Thomson Reuters found them hallucinating in one out of six or more benchmark queries. In 2023, two New York lawyers filed a brief with citations to six nonexistent cases. That was Mata v. Avianca, and it was not the last. As of May 2026, a public database maintained by researcher Damien Charlotin listed 1,397 U.S. cases in which lawyers or self-represented litigants filed documents with hallucinated content, and courts have imposed monetary sanctions on lawyers in many of them. No doubt the number is much higher today. The threat has attorneys skipping AI, spending extra time proofreading, delegating reviews, and double-checking.


Confidentiality, privacy, and privilege

Using AI exposes attorneys who must adhere to high standards, including the rules of professional conduct. Business confidential information about the law firm and about clients has to be protected. Some AI applications retain what you type and share inputs with third parties, so reading the terms is not optional. Stanford's catalog of legal AI risks includes inadvertent disclosure of confidential information and waiver of privilege. In Law360 Pulse's 2025 survey, client confidentiality, security, and data privacy ranked right behind AI's imperfect grasp of legal ethics among lawyers' concerns. North Carolina's 2024 Formal Ethics Opinion 1 says a lawyer may use AI, but must do so competently, securely, and with independent judgment in supervising the work.

Ethics, liability, and who is accountable

The rules haven't changed, but the questions have. ABA Formal Opinion 512 applies the existing rules, including competence, confidentiality, communication, fees, and supervision, to generative AI. It doesn't create new ones. Malpractice claims ranked second among firm lawyers' concerns in Law360's 2026 survey.

Who is accountable? Lawyers must supervise the nonlawyers who assist them. AI sits in a gray zone as to whether it counts as one. A vendor builds the tool, the firm configures it, and a lawyer signs the work. Regardless of how responsibility is delegated, the lawyer's name is on the line.


Pipeline and apprenticeship

Firm lawyers in the Law360 survey also worried about fewer opportunities for newly minted attorneys to do basic legal work. That work is how lawyers learn. If AI does the first pass, the learning path for inexperienced attorneys narrows, and the career pipeline can choke. No more apprenticeship.


Billable hours and revenue

Bloomberg Law notes that when rote work takes seconds, clients may demand alternative billing structures, which could consolidate the number of inexperienced staffers. Commentators have speculated that AI is the harbinger of death for the billable hour, and one piece this summer concluded it isn't dead but AI is chipping away at it. Either way, the disruption of time-honored revenue models cuts deeply into a hierarchical culture of tenure and king rainmaker.

ABA Formal Opinion 512 is clear that lawyers charging hourly may bill only for the time actually spent. If AI helps an attorney reduce a task from three hours to one, the attorney can only charge for one hour. It's not effective to bolt an AI solution onto existing financial and pricing models. Redesign is needed.


Trust gap and extreme uncertainty

Adoption is outrunning trust. Filevine's 2026 AI Trust Index, a vendor survey of 115 legal professionals, found 67% using AI weekly, yet only 3% extremely confident in it, and just 1% extremely confident for legal work. In Law360's survey, firm lawyers using AI three or more times a week were less positive in 2026 than in 2025: 51% versus 73%. The more people use it, the more realistic they get.

Then there are the unknown unknowns. AI tools change so rapidly that it’s difficult to understand how they work. With agentic AI developments, some AI tools now act autonomously, integrating with systems and completing tasks in unexpected ways.

The middle ranks

Inherently troubling for the legal industry are injustices and inequities: models trained on biased data can produce discriminatory outputs, as Stanford and Bloomberg Law both note. There's unauthorized practice of law, which a 2026 scholarly paper on agentic AI lists among the professional duties these systems put at stake. And there's job loss. In Law360's survey, 29% of firm attorneys worried about AI taking white-collar jobs, compared with 54% of in-house attorneys.

Bringing up the rear

At the back of the parade are fears no less important, but more general about AI use. Legal professionals and leaders in other industries keep having real conversations about token cost, environmental impact, affordability, decreased access to justice, and inequalities in representation.

Introductory prices and free versions of AI tools have been the norm for several years. As the models become more capable and investors keep pouring money into AI products, many wonder when the bubble will burst and what it will cost consumers who have grown accustomed to practicing law with AI assistance. Others wonder whether fewer clients will be able to pay. Even if these concerns seem like stragglers, water, land, and quiet near data centers are also on people's minds. And access to justice cuts both ways. We can't name every fear, which makes the parade feel endless.



A named fear can be managed

Again, I'm generally overwhelmed even though I’ve been told that one of my superpowers is the ability to stay calm in the midst of chaos. Yes, my coping skills are on full display. Want to know the secret? Named fears can be overcome. Sound simple? It definitely is not complicated. Let me explain.

Fears are indicators. Name the fear, map it to a risk, mitigate it, and manage it. Each of the fears is tied to a risk that has a control:

  • Hallucinations: verify every authority before it leaves the building, use legal-specific tools, and write down who checks what.

  • Confidentiality: read the terms, use enterprise agreements, and keep client information out of consumer tools.

  • Ethics and accountability: adopt a written AI use policy, train everyone, and monitor compliance.

  • Pipeline: redesign apprenticeship so everyone learns through mentoring and reverse mentoring by checking and improving AI work.

  • Billable hours: explore flat and alternative fees that reward outcomes, not minutes.

  • Trust gap: start with low-risk uses, document results, and expand as trust is earned.

The controls are easy to say, difficult to document, and hardest to do.


Fear factor

How much fear can you tolerate? A simple risk matrix or assessment helps in understanding how dangerous each AI use is, rating how likely something is to go wrong and how bad it would be. When you align with practices to assign an owner,  attach a control, and revisit the risk on a schedule, lawyers have a defensible, risk-based way to governing AI. The goal isn't zero risk. The goal is a reasoned, documented process attorneys can share with a client, a judge, or an insurer. This work is how Rivers Strategic Advisory helps law firms.


Fearful and optimistic

As lawyers, we need to use AI with competence and in a way that bolsters the trust we've established as an elite profession. Fear around using AI is real, and I'd question anyone who doesn't at least feel cautious. It's scary. If you’re scared, say you’re scared. Interrogate the feeling and reframe the emotion. We've lived through scary before. Every generation of lawyers has faced a technology that disrupted the way we work, and the profession adapted and thrived. Using AI is a real opportunity to innovate legal culture, transform the legal industry, and help people.


If you enjoyed this read, sign up to our email list to get the blog directly in your inbox.



Sources

 

 
 
bottom of page