This spring, the California State Bar proposed amendments to Rules of Professional Conduct that would require lawyers to independently verify every AI output used in client representation, disclose AI use when it materially affects the scope or cost of representation, and scrutinize the data practices of every AI tool they use.

As lawyer and former National Law Journal editor-in-chief Bob Ambrogi wrote in his coverage of the proposals, they represent among the most detailed and comprehensive AI-specific rule amendments any state bar has put forward. The ABA’s Amanda Robert echoed this point in her own analysis of the first-time proposed amendments.

The proposals are a response to a specific problem. AI tools are now widely available, widely adopted, and outpacing the profession's ability to govern them. But embedded in that regulatory challenge is a more fundamental question that the rules can't fully answer: If everyone has access to AI, what really makes an organization, a firm, your team, stand out?

The answer isn't the technology. It's hiring patient, communicative, well-rounded, and flexible people to use it.

The Research Isn’t New

The case for prioritizing soft skills in hiring isn’t new. It's just been consistently underestimated. Research conducted by Harvard University, the Carnegie Foundation, and Stanford Research Center concluded that 85% of job success comes from having well-developed soft and people skills, with only 15% attributable to technical knowledge. A separate study of Fortune 500 CEOs found that 75% of long-term job success depends on people skills, while only 25% depends on technical ability.

None of this is abstract. In a world where AI tools are rapidly closing the hard-skills gap, where any attorney can run a legal research query, any paralegal can generate a first draft, and any associate can summarize a deposition in seconds, the qualities that can't be automated become the real differentiator. Adaptability. Curiosity. Emotional intelligence. The ability to ask the right questions.

The Human Element in the Tech Era

There's a question that doesn't get asked often enough in conversations about AI: when everyone has the same access to technology, what is it that cannot be automated?

The answer, over and over again, comes back to human judgment applied to context. Law students using AI for research aren’t differentiated by their ability to run queries. They're differentiated by their ability to know which questions are worth asking, to recognize when an answer is incomplete, and to understand the human stakes of the legal issue at hand. AI surfaces information, but critical thinkers interpret it.

The law firms and companies pulling ahead aren't just adopting AI. They’re coupling it with individuals who know how to leverage the resources available to them. Those people don't always come from the expected backgrounds. They're curious generalists as much as narrow specialists. They're the ones who will catch a hallucinated citation, push back on an AI summary that misses the point, and ask a follow-up that the tool didn't think to prompt.

At Steno, this shapes how we think about hiring. Our team looks for people who are engaged with the world, who pursue things out of genuine interest and bring a life outside their job description to the table. People like that tend to adapt faster, collaborate better, and approach novel problems with more creativity. When the technical playing field flattens, those are the qualities that don't.

What California's Proposal Actually Says

The proposed California amendments initiated by a directive from the State Supreme Court itself. In an August 2025 letter to the State Bar, the court asked its professional responsibility committee to consider making the bar's “2023 Practical Guidance for the Use of Generative Artificial Intelligence in the Practice of Law" into enforceable rules, and to address agentic AI specifically. The public comment period closed in early May. The rulemaking process is ongoing.

What makes California's approach notable is not just the content of the proposals, but the mechanism. Most states have addressed AI through ethics opinions, which carry persuasive authority but not disciplinary force. California is proposing to embed these requirements in the enforceable Rules of Professional Conduct.

The line running through the proposed amendments is straightforward. Lawyers must independently review and verify every AI output used in client representation. They must disclose AI use when it materially affects the scope, cost, or decision-making process of representation. They must ensure confidential client data isn't being exposed. They must verify that every cited authority, including AI-generated citations, actually exists and says what it's claimed to say. And firm leadership must establish actual, functioning AI governance policies that apply to all staff.

The Rules’ Limits

It's valid to question whether this proposal runs counter to technological progress. The core attraction of legal AI is speed. A blanket "verify every output" requirement structurally doubles the work—AI produces, lawyer re-verifies— no tiered framework to distinguish routine tasks from high-stakes ones. Agentic AI, systems that can plan and execute multi-step tasks with minimal human intervention, is the direction the industry is moving. The more autonomous the workflow, the harder compliance becomes.

This isn't to say the impulse behind the rules is wrong. Courts have sanctioned lawyers for submitting AI-hallucinated citations. Confidential client data has been exposed through careless use of AI tools, emphasizing the importance of legal AI frameworks. But whether uniform caution applied regardless of context or stakes is the right balance is exactly the kind of question that requires human judgment to answer. No AI tool is going to inherently resolve it.

What the proposed rules are actually describing, underneath the compliance language, is a set of tasks that can't be delegated to the technology itself: catching what the tool got wrong, knowing when to push back, understanding the stakes well enough to know which outputs need scrutiny. Those aren't technical skills. They're the same qualities, like curiosity, judgment, and professional grounding, that separate the people who use AI well from the ones who don't. California is trying to mandate in regulation what good hiring would produce naturally.

The firms and companies that will do best in this environment aren't just the ones that adopt AI fastest. They're the ones that hire people who know when to trust it, when to push back on it, and when to ask a question it didn't think to raise. Those people exist. Hiring for them takes intentionality.

Book a demo to discover how Steno can support your team.