Sustainable Intelligence

The question is not whether to use AI. It's how to use it with intention.  

Over the past few months, I've had the opportunity to present on this topic through two continuing legal education programs - one for a prosecutor's office and one for a global law firm. The audiences couldn't have been more different yet the AI concerns were nearly identical.

That made me wonder how many others are facing the same challenge: not just how to use AI competently, but how to do it in a way that doesn't erode the capacity that competent judgment depends on.

AI as a Risk Multiplier 

AI adoption in legal work is moving fast. Roughly three-quarters of corporate legal professionals now use generative AI at least once a week, and a similar proportion of law firm lawyers expect it to be central to their workflow within five years.

That adoption is reasonable. AI can assist lawyers in a number of ways including streamlining drafting and contract review, providing predictive analysis, and offering trial preparation support. The question isn't whether to use it. The question is what happens when you use it inside a system that's already overloaded.

An Overloaded System

The disciplinary record is already reflecting system overload. In Oregon in December 2025, an attorney faced $110,000 in sanctions and case dismissal after 15 references to nonexistent cases and 8 fabricated quotations appeared across three briefs. In Massachusetts in April 2025, an attorney received a public reprimand and a $2,000 sanction for failing to supervise an associate who submitted AI-generated citations without verification.

The human story in both cases is insufficient review, likely under conditions of time pressure and overload.

Most of us would like to think we would never find ourselves in the position of those lawyers. But under pressure and cognitive overload, the human brain reliably reaches for shortcuts that may lead to trusting tools that appear authoritative or accepting drafts without sufficient analysis.

Predictable Responses to Overload

Automation Bias

One predictable response to cognitive overload is automation bias. It's the tendency to over-rely on automated recommendations, even when contradictory and more accurate information is available. 

The problem isn't that AI makes mistakes. The problem is that humans stop catching them - and catching them takes a kind of attention that's hard to sustain when you're already running on empty.

Pattern Recognition

Another predictable response is pattern recognition. Humans use pattern recognition as a cognitive shortcut, the process of cognitively organizing familiar inputs so we don't have to start from scratch every time.

Used with awareness, pattern recognition helps form the basis of expertise. Think of the experienced litigator who reads a familiar judge's demeanor and pivots her approach in real time. That's pattern recognition, and it is valuable. But that same judgment made without fresh analysis, or under overload, carries risk.

Uneven Ground

Digital pressure and AI adoption don't land the same way for everyone, and that gap matters for wellbeing as much as for fairness. Women lawyers report significantly higher rates of burnout and attrition. Attorneys aged 25 to 34 are two to four times more likely than older colleagues to report moderate or high stress. Lawyers from underrepresented groups make up 30% of associates but less than 10% of equity partners, and the non-promotable tasks that drive overload and over-functioning fall disproportionately on women and people of color.

That matters because depletion and verification are connected. The lawyers carrying the heaviest overload are often the ones with the least bandwidth to slow down and check AI output carefully — which means the risk of automation bias isn't distributed evenly either. AI tools trained on historical data can also reproduce existing bias in their output. Reviewing AI work for accuracy matters. Reviewing it for whether it disadvantages certain voices, arguments, or perspectives is part of the same responsibility.

Four Strategies To Protect Your Capacity

• Protect attention deliberately. Focus work such as legal analysis, AI output review, and ethical decisions cannot tolerate interruption the way routine tasks can. Scheduling even one protected 60 to 90-minute block per week and batching administrative tasks are small structural changes that can have measurable impact. 
• Build AI verification habits. Have a verification system for all authority – every citation, every case, every source. Check whether a summary accurately reflects the actual document and applies to your jurisdiction and specific facts. A useful test: if you had to defend this in front of a partner or a court, could you? 
• Get clear on boundaries to manage overload. Before responding to any request, ask: What's actually urgent? What can wait? What level of review does this require? 
• Protect your intuition. Legal training teaches us to distrust what we can't cite. But experienced lawyers know that the feeling something is off often precedes the evidence. The signal a client isn't telling the full story, the nudge to ask one more question, the sense a colleague needs a check-in before burnout becomes visible – those are human response signals that AI cannot generate. Before a high-stakes conversation or decision, take 30 seconds and ask yourself: is this a moment that calls for my own judgment rather than AI's output? Notice what your body is telling you.

Sustainable Intelligence 

AI does not have to be the enemy or the crutch. The key is using it with awareness and giving yourself permission to not use it in every area of your work. That is not a failure of efficiency. And protecting your judgment is not separate from protecting your wellbeing. That's sustainable intelligence.