A practical framework for role-appropriate AI literacy, durable skills, and responsible governance in legal work

[EDRM Editor’s Note: The opinions and positions are those of the author.]

Introduction

AI is already part of legal work. The real issue is that many teams are still asking the wrong question.

The question is not whether everyone needs AI skills. It is what kind of understanding each role now requires, and what happens when that understanding is missing.

In legal practice, the risk is not just inefficiency. It is relying on outputs that appear sound but cannot be explained, tested, or defended when it matters. Speed can create confidence. That confidence is not always earned.

Not everyone needs to become a technical expert. But everyone now needs a working level of AI literacy. That means understanding how AI affects their work, what risks it introduces, when outputs are useful, and when professional judgment must take over. It also means knowing when a process will hold up under scrutiny and when it will not.

AI exposure is uneven across roles, and exposure is not the same as replacement. Outcomes will depend on the work itself, the skills professionals bring, the training they receive, and the judgment applied at each step.

This is not just a technology shift. It is a professional shift, and it is already underway.

Curiosity and AI

The legal profession has always rewarded curiosity. Strong legal professionals do not accept information at face value. They ask what may be missing, what assumptions need testing, what facts do not fit the initial story, and what else might explain the record. AI can amplify that instinct, but only if the user pushes beyond the first answer.

Cat Casey frames this moment as one that begins with curiosity rather than fear:

AI literacy starts with curiosity, not fear. The legal professionals who will thrive are the ones willing to play with the tools, test the edges, ask sharper questions, and build new habits before someone else defines the future for them. That is what we are bringing together at Masters AI: an engaging, dynamic space where legal teams can move from AI anxiety to AI fluency without losing the judgment, skepticism, and precision that make legal work powerful.

Cat Casey, Founder and CEO, The Technocat LLC.

That distinction matters. AI fluency does not replace legal judgment. It means learning how to use new tools while preserving the skepticism, precision, and discipline that legal work requires.

A basic prompt might ask an AI system to summarize a set of documents. A more effective approach asks what is missing, where the gaps are, which sources have not been explored, what changed over time, and why one timeline conflicts with another. AI may surface a pattern, but curiosity turns that pattern into an investigation.

In eDiscovery, that difference is critical. During early case assessment, AI might flag a spike in communications during a key period. That is useful. But the real work begins after the pattern is identified. Does it reflect ordinary activity, or something else?

AI identifies signals. Curiosity helps interpret them.

Curiosity surfaces the right questions. Critical thinking evaluates the answers.

Critical Thinking and AI

One of the most common mistakes in AI use is equating speed with accuracy. They are not the same.

AI can process large volumes of information quickly. It can summarize, classify, and prioritize. Those capabilities improve workflows. They do not guarantee correctness.

Plausible does not mean reliable.

That distinction matters in legal work. Outputs may appear complete while omitting key facts, mischaracterizing context, or overstating conclusions. Without verification, those risks go unnoticed.

AI should be treated as decision support, not decision replacement.

Gina Taranto emphasizes the importance of asking questions. That willingness to question, validate, and admit uncertainty is central to responsible AI adoption:

Some of the most important skills all of us in legal technology can flex right now involve getting comfortable asking a lot of questions, answering with an “I don’t know” when that is the truth, and working with a team to get answers that allow us to apply critical thinking, exercise judgment, and continue developing individual and collective AI literacy.

Gina Taranto, PhD, Executive in Residence for Language Technology and AI Innovation, EDRM.

Effective use requires discipline. Outputs must be tested against the underlying record. Assumptions must be challenged. Gaps must be identified. What is missing often matters more than what is returned.

In practice, this shows up in review workflows and early case assessment. A generated summary may streamline analysis, but reliance without validation introduces risk. Context can be lost. Meaning can shift.

Confidence scores do not replace judgment.

AI assisted platforms can identify relevant documents and prioritize review. That has value. Professionals remain responsible for determining whether results align with case strategy, whether privilege is protected, and whether conclusions are defensible.

Critical thinking is not reduced by AI. It becomes more important.

Communication and AI

Legal work depends on helping others understand information well enough to act. AI can generate content. It does not ensure clarity, judgment, or persuasion.

More output does not mean better communication.

In practice, the same facts must be translated differently for different audiences. Outside counsel may require defensibility. Corporate legal may focus on risk and cost. IT expects technical precision. Executives need clear direction.

AI can assist with drafting. Professionals control the message.

That control goes beyond tone. It requires deliberate choices about structure, emphasis, and risk framing. Outputs must be reviewed, refined, and aligned to purpose before they are relied upon.

AI also changes expectations around trust. It is no longer enough to present conclusions. Professionals must be able to explain how outputs were generated, what assumptions were made, and where limitations exist.

Clarity builds understanding. Transparency builds trust. Both remain the responsibility of the professional.

Persuasion and AI

Persuasion moves work from insight to action. AI can support analysis, but it does not create buy in.

Buy in comes from understanding concerns, framing the issue clearly, and connecting recommendations to practical realities.

This is especially important when AI adoption raises questions about risk or role impact. Teams want to understand not just efficiency, but governance, oversight, and accountability. AI can support a proposal. Persuasion helps it land.

Addressing questions about defensibility, oversight, and risk before they are raised is what builds trust.

Adaptability and AI

Legal work has already been shaped by decades of technology change. AI is the next acceleration.

Adaptability is not about chasing tools. It is about staying engaged, applying judgment, and translating existing expertise into new value.

A litigation support professional who uses AI to improve workflows or documentation can become a resource for implementation and governance. The role evolves.

Teaching and AI

AI can support learning, but people build learning cultures.

Professionals can use AI to develop training materials and documentation more efficiently, but mentorship and shared experience provide the context and judgment tools cannot replicate.

Judgment and AI

Judgment remains central. AI can suggest, rank, and prioritize. It cannot assume responsibility.

In legal work, that distinction matters. Professionals still decide what is produced, what is privileged, and what obligations apply.

AI supports the process. Judgment determines the outcome.

AI can accelerate decisions. It cannot assume responsibility for them.

Governance and AI

As organizations adopt AI, the central question is not what the technology can do. It is what information it is being given.

Before using AI, the first question must be: Where is the data coming from?

Legal and business information is often dispersed across multiple systems. Its use in AI requires clear understanding and disciplined control. At a minimum, professionals must determine:

• Where the data resides
• Who controls it
• Who has access
• Whether it includes confidential or privileged information
• What legal or contractual obligations apply
• Whether the tool is approved for use
• How the provider handles, stores, and retains the data
• Whether the process is transparent, explainable, and defensible

The risk is not confined to the tool. The greater risk is failing to understand the data placed into it.

Uploading draft agreements, internal strategy documents, or investigation materials into an unapproved system introduces exposure where data handling is uncertain or uncontrolled. That risk cannot be mitigated after the fact. Processes must not only be defensible, but documented in a way that allows them to be explained and validated after the fact.

The legal framework is still developing. The extent to which AI use is protected by attorney client privilege or work product doctrine remains unsettled. Courts, regulators, and ethics authorities continue to assess how existing protections apply when sensitive information is shared with or processed by third party systems. There is no consistent or reliable standard.

This is not a technical issue. It is a governance issue.

AI literacy therefore requires more than understanding outputs. It requires disciplined data awareness and judgment. Before asking what AI can do, professionals must determine whether the data should be used at all.

AI does not reduce responsibility. It increases it, particularly where decisions are accelerated and applied at scale.

Practice Points for Legal Teams

• What problem are we solving
• What data is involved
• Where does that data reside
• Is the tool approved
• What obligations apply
• How will outputs be reviewed
• Who is responsible for decisions
• What training is required
• How will the process adapt
• Can the process be defended

These questions do not slow progress. They make it sustainable.

The Real Competitive Advantage

As AI advances, human skills become more important, not less.

AI can support execution. It cannot replace curiosity, critical thinking, communication, persuasion, adaptability, teaching, judgment, or governance.

The legal profession has always relied on the balance of people, process, and technology. AI reinforces that balance.

The future will be defined by professionals who use AI to ask better questions, make sound decisions, communicate clearly, and manage information responsibly.

AI can accelerate the work. Judgment determines whether it stands up when it matters.

Opportunity Notes

Continue the Conversation at Virtual Lunch with Leaders

Every Friday at noon Pacific, a new conversation emerges at the San Diego Paralegal Association’s Virtual Lunch with Leaders (SDPA VLL). This Friday, June 12, VLL’s topic will be this article. There is no charge for this session, and all are welcome. Register to attend: Virtual Lunch with Leaders – Curiosity + AI – June 12 | 12 PM PT

Continue the Conversation at Masters AI Legal LA, with CA CLE

For legal professionals wondering where to begin, one answer is simple: get in the room, ask better questions, and stay curious.

On June 17, Masters AI Legal brings together Am Law partners, Fortune 500 legal leaders, academia, legal tech builders, and the professionals accountable for how AI actually lands in legal practice. Hosted at Buchalter in Los Angeles, with virtual access available for the first time, the program offers legal professionals an opportunity to move beyond abstract conversations about AI and into practical discussions about literacy, judgment, governance, ethics, and real-world implementation.

This program offers California Continuing Legal Education (CLE) credit provided by the San Diego Paralegal Association (SDPA) in association with the Electronic Discovery Reference Model (EDRM). Attendees can earn six CLE hours, including one hour of ethics credit. Register to attend: Masters AI Los Angeles 2026 – June 17

Assisted by GAI and LLM Technologies per EDRM’s GAI and LLM Policy.