Becoming an Associate in the Age of AI
By A fellow attorney friend told me about the moment when one of her clients paused in a meeting to consult ChatGPT regarding their legal approach. She described hearing the client’s nails clacking on a keyboard. Her supervising partner started to say something, stopped and then asked:
“Are you using ChatGPT by any chance?”
The client said, “Yes, and it says right here that your strategy to . . .”
I don’t really need to finish that story, do I? Because in 2026, I would put money down that many attorneys have had a similar — if not identical — interaction.
Since that watershed moment in November of 2022 when OpenAI released ChatGPT for public consumption, generative AI has become an all-in-one pocket doctor, therapist, lawyer and priest for many consumers. So, what are we to do, especially if we are new attorneys, in this age of AI?
Jump on the Bandwagon, or . . . Perish?
The arguments for and against using artificial intelligence in one’s practice have been repeated ad nauseam. Most of us have heard one or more variations of the following:
“Those who don’t learn about artificial intelligence and adopt it into their practice as soon as possible will be left behind.”
“If you use AI, you’re buying a one-way ticket to malpractice.”
“I love ChatGPT. I use it for everything!”
“I refuse to touch anything related to AI.”
And unfortunately, sometimes, humans can find it difficult to approach an issue with nuances in mind. It’s possible to learn about AI and adopt it at a reasonable pace into our practices without forgetting how to be an attorney on our own. But how can we do this?
I think it’s time to bring back a word and practice that’s lost popularity: discernment. By discernment, I refer to the practice of detecting and recognizing one thing from another. The word is often used in a way that connotes taking one’s time to discern. It tracks that this term has a downward trajectory in usage considering the increase in deployment of convenience technology.
Everyone feels rushed these days, especially attorneys, although the general public might disagree, but I digress. And for those who do want to take their time, it’s implied that they must be trying to pad the bills or some other nefarious reason as opposed to simply wanting to be thorough.
What do we do? How do we discern when to use artificial intelligence in our work and when not to, especially considering the various security, privacy and environmental concerns surrounding using large language models like ChatGPT, for example?
Below is a list of questions or checkpoints that we can use to discern whether to incorporate artificial intelligence tools into our work lives.
Task Discernment
- Have I done this task before? Have I mastered it on my own?
- Is there an impending timeline for this task?
- Do I understand the importance of this task for the client?
- Would using AI on this task help the client?
- Is this a rote task well-suited for automation?
- Am I using this for substantive legal work, administrative tasks or marketing work, such as drafting social media posts?
- Will using AI help me effectively and correctly finish this task more quickly?
Ethics Discernment
- Does this tool that I am using align with our Code of Ethics as attorneys?
- Does it protect client information and employ proper security protocols?
- Has this tool been tested effectively? Is it new? Have I reviewed how it has been trained and tested and on what data?
- What data bias may be affecting its output, and have I considered how it could influence the client’s case, if at all?
- Have I refrained from inputting client information?
- Have I fact- or cite-checked the output?
- Does using AI for this particular use case feel acceptable in light of my personal values or moral code?
If you answered “yes” to many of the above questions, then using AI may be a good idea. If you answered “no” to many of the above questions, then foregoing AI and completing “manually” might be better.
The key here is to be open to using new tools but not at the expense of your intuition and values. If there’s a part of you that doubts employing AI for a particular use case, don’t use it. You can always try again in the future if and when you’re able to establish more confidence in the tool. Frankly, not everything needs to be automated.
A Note on Agency and Ethics
If I am being candid, the thought of AI makes me sad. Living in a world where people want to deploy AI-bots to create art, music, film and poetry, among other things, has an uncomfortably dystopian feel to it. After all, creativity and the act of creating is one of the most beautiful, redeeming qualities of humankind. Personally, I think it is a natural response to any historic technological development — grieving the loss of “what once was” is okay. To bury our heads in the sand is not. Especially as attorneys who have taken oaths to maintain a certain standard of professional duty.
Facing the reality of technological developments is one such duty, which is why I truly believe that taking a beat to discern when, how and why we use AI is critical. The reality is that anyone using AI tools in this moment is still a part of a very small group of people in the world. “Global adoption of generative AI tools reached 16.3 percent of the world’s population, up from 15.1 percent in the first half of 2025…”
The people leading the pack in rapid AI adoption may sound convincing in saying that we need to use AI for everything, everywhere and all at once. Practically speaking, though, this is not the case. We also need people who take their time to ask questions and discern.
At the end of the day, our agency is all we have — agency over how we operate in the world and how we choose to evolve as individuals. I don’t mind using AI. But I also don’t want it to become my crutch. I deserve to try things out for the first time so I can understand them on my own. I deserve mentorship on topics without being directed immediately to chatbots. I deserve to desire human communication and connection in my work and personal life.
You also deserve these things. And most importantly for the purposes of this article, so do our clients. Yes, we should be open to finding ways to adopt technology that would benefit our clientele either substantively in how we strategize legally or financially if it means making our services more economical. But adopting technology for the sake of “being the first and the fastest” may undercut those efforts too.
This is why discernment is necessary. And I hope this little blog post can help you start that process of creating a checklist of when and how to use AI in your life.
This Blog Post: A Use Case
I figured it would be fitting to use this post as an example of how AI LLMs can be used without completely getting rid of our agency. To start, this post started as an idea that I have struggled with: the when, how and why of using AI as an associate.
Instead of starting with AI, I brain dumped my thoughts into an initial, very rough draft. There was no time constraint for this, so there was a lot of the good ole “sit back, stare at the wall, walk away” type of thinking as I tried to put into words what I wanted to say on this matter.
But then, after a brief discussion with a connection about all things AI and agency, I thought: “why not use this as a use-case study and plug my draft into an LLM?” Once my draft was done, I added it into Claude with the following prompt: “This is a rough draft of an [article] I want to submit to the NCBA for their blog. Please review for clarity, organization, and grammatical errors.”
Claude responded with a list of suggestions as I didn’t want to copy and paste. My voice is important to me as a writer, so therefore, I don’t want Claude to rewrite but rather provide feedback. How you prompt an LLM is incredibly important for this reason.
Funnily enough, I didn’t agree with some of the feedback. For example, one of the responses was “The sentence ‘Living in a world where so many people want to deploy AI-bots to create art . . .’ runs long and loses its punch. Consider breaking it up.”
At first, my reaction was “No. I don’t want to change it. I like the way it sounds.” After a round of edits with the NCBA team, I ultimately came to the same conclusion: the sentence needed work.*
I stand by my belief that variations in writing styles are what makes writing fun. Humans don’t always talk in perfectly written sentences suited for newspaper reporting. Maybe Hemingway did, but occasionally, a long sentence is OK if it proves your point. But just because AI says so doesn’t make it gospel.
I then had a good friend read my draft for clarity and feedback (Thanks, Amanda!). After Amanda read it, I re-read it several times and finally read it out loud, per Bob Woodward’s advice that I heard at a college journalism conference back in 2017. Once somewhat satisfied, I submitted.
This is just one way that we can use AI without completely getting rid of our agency. Ironically, the key to this is pausing and discerning. In a world where many businesses and tech leaders want you to “move fast and break things,” this will seem countercultural. Yet these principles are key if we want to avoid a race to the bottom.
*We humans are also stubborn. Especially writers.