You Cannot Wait This One Out
AI is here and it is permanent. The lawyers who win this era are the ones who get fluent now while everyone else is still hoping for normal to come back. Normal is not coming back. But winning is absolutely still on the table.
The Cat Signal · Issue #1 · Week of May 12, 2026
The Signal
Buckle up because this is not slowing down.
Most legal AI content is written for the same ten thousand lawyers. The ones already in the cool kids tech club. The ones with innovation budgets and Chief AI Officers and dedicated legal tech teams. You know the ones.
There are 1.3 million practicing lawyers in this country. The other 1.29 million are navigating this transition largely on their own. Watching from the outside.
Trying to figure out what is real, what is hype and what they actually need to do about it before the window closes.
This is for them. This is for you.
Here is the thing about AI that nobody wants to say plainly. AI is here and it is permanent. Thanos already snapped. The legal profession was not spared and the bubble is not going to burst. Waiting it out is a strategy that has already cost people their edge. The lawyers who win this era are the ones who get fluent now while everyone else is still hoping for normal to come back.
Normal is not coming back. But winning is absolutely still on the table.
I am Cat Casey. Chief Legal AI Futurist at Masters AI and two decades deep in legal technology. Gibson Dunn. PwC. Reveal. I wrote the Wiley book on AI in legal tech. I keynote the rooms where this gets decided. And I have spent years watching brilliant lawyers get left behind simply because they did not have the right resource in their corner at the right time.
The Cat Signal exists to fix that. Every week I bring you what is actually moving in legal AI, what I am hearing on the main stage, what the community is debating and the one idea worth your full attention. No jargon for jargon’s sake. No insider speak. Just a straight line from where legal AI is to where you need to be.
Masters AI is your front page for all of it. Eleven cities. A community of practitioners doing the actual work. The Signal right here every week. You do not have to figure this out alone and you do not have to be in the cool kids club to get fluent.
You just have to show up. You already did.
The Docket
Item 1 — California just proposed requiring lawyers to verify every single AI output. Every one.
The California Bar’s COPRAC approved proposed rule amendments in March 2026 that would formally require lawyers to independently review and verify any AI generated output before it goes anywhere near a client or a court. The California Supreme Court itself set this in motion, directing the Bar to address general AI use and specifically agentic AI tools — the systems that research, draft and revise with minimal human supervision. Six existing rules get rewritten. Competence, confidentiality, candor toward tribunals, supervision. All of it.
Why it matters: California moves first and the rest of the country follows. If you are using Harvey, CoCounsel or any AI tool in your workflow right now, this is your preview of what your bar will be asking of you next.
Item 2 — Federal judges are on record. Lawyers should be paying attention.
At the IAPP Global Summit in Washington DC last month, Chief Judge James Boasberg of the DC District Court and Judge Allison Burroughs of the Massachusetts District Court went public on AI in the courtroom. Boasberg described presiding over filings packed with hallucinated citations and sanctioning the attorneys responsible. Burroughs put it plainly: lawyers can use AI but they must disclose it and they must verify every citation. Worth noting the AI features inside the research tools judges themselves use have been largely blacked out while courts figure out the rules.
Why it matters: The people who sign off on sanctions just told you exactly what they expect. This is the bench speaking directly to the profession. Listen accordingly.
Item 3 — 79% of legal professionals use AI. 53% of firms have zero governance policy.
The Clio 2025 Legal Trends Report confirmed what a lot of managing partners do not want to say out loud. Nearly four in five legal professionals are already using AI tools in their work. More than half of their firms have no formal policy governing that use or staff have no idea one exists. The gap between adoption and oversight is exactly where sanctions happen, where confidentiality breaches happen and where clients start asking questions firms are not prepared to answer.
Why it matters: Shadow AI — lawyers using consumer tools like ChatGPT without IT visibility — is not a future risk. It is already running inside your firm. A guardrails policy is essential anymore and California just made that official.
Main Stage Musings
Chip Koons · Digital Forensics Expert
AI just changed what counts as evidence. Most lawyers have not caught up.
Chip Koons laid out a reality check the room needed. Investigators used to follow the footprints of human behavior through link files and local drives. Now they follow prompts, chatbot histories and AI-generated drafts — and if you do not know what questions to ask, you are missing evidence that could win or lose the case. His concept of the “liar’s dividend” is the one that sticks: AI manufactures fake evidence and casts doubt on real evidence at the same time. Courts are still working out whether prompts are discoverable, whether they are privileged and how you authenticate an AI-generated document for a judge who has never seen one. Fortis Advisors v. Crafton is the early signal. It will not be the last.
The room takeaway: the legal community can wait for problems to define the guardrails, or it can get ahead of them. The window is narrowing fast.
Martin Tully · Partner, Redgrave LLP
Agentic AI goes beyond answering questions. It takes action. That changes everything.
Martin Tully opened by invoking the 2021 Zoom hearing where a lawyer pleaded with a judge that he was, in fact, human — not a cat. Five years later his point was sharper: in 2026 the cat might actually be your co-counsel. Agentic AI pursues goals, retrieves documents, synthesizes conclusions and — if you have given it the permissions — emails the result to opposing counsel at 2 a.m. His governance framework came down to four verbs: see, infer, write and send. Most organizations have not made deliberate decisions about any of them. And his warning on blanket bans was blunt: bans push adoption underground and strip firms of visibility. Ungoverned AI keeps happening — it just moves outside your visibility.
The room takeaway: the permissions architecture matters more than the interface. Who trained it, who feeds it data and who is responsible when it knocks the vase off the table — that is the whole game.
Matthew Kerbis · Attorney and Podcaster
AI raises all boats. That is exactly the problem.
Matthew Kerbis made the counterintuitive argument of the day. The real threat AI poses is something more subtle than raw power — it is that it is making everyone sound the same. Spelling errors gone, passive voice gone, but when every professional runs through the same tools the output converges. Everyone has B-minus writing now. His answer is NotebookLM, Google’s source-grounded platform that answers only from sources you provide — contracts, briefs, CLE presentations — with no hallucinations from training data. Law firms can build shareable client-facing notebooks that deliver expert answers on demand without billing by the hour. Taste, he argued, is essential. That is the differentiator AI cannot replicate.
The room takeaway: curation and subject matter expertise are the competitive edge. The professionals who win are the ones who bring genuine judgment to how they use these tools.
Word on the Street
The conversation the legal community keeps having in hallways and DMs but not yet saying loudly enough: AI literacy is the new wealth gap. When I wrote my latest Forbes piece on the costliest mistakes of the first industrialists, 117,966 jobs had already been cut and attributed to AI-driven restructuring since January. That number is materially higher today. We are watching a fire sale on human capital dressed up as transformation language. Carnegie built his libraries after the funerals. We still have time to build them before.
For lawyers this is not abstract. AI fluency is already separating the lawyers who are expanding their practices from the ones quietly losing ground. The gap is about who decided to get fluent and when. The right moment was six months ago. The second best time is now — and the community showing up to do that work is exactly what Masters AI exists for.
Next up: Masters AI Los Angeles · June 17 · Join us →
Warmly,
Cat Casey
Chief Legal AI Futurist · Masters AI · TechnoCat