About | AI for Legal Institute

mission

Almost everything a firm reads about legal AI was written by someone selling it.

Law firm leadership is being asked to make durable decisions — what to adopt, who checks the output, what to charge, what happens to client confidences — on the basis of material produced by vendors, consultants selling implementations, and commentators predicting transformation. Very little of it is evidence, and almost none of it is disinterested.

The AI for Legal Institute exists to be the part that is not selling anything, and to make the first hour of understanding free. Any firm can use the Institute’s adoption concierge at no charge: it asks where the firm is, what it is trying to decide, and what is already being used without approval; it explains what the decision actually turns on and what other firms have found; and it says plainly which questions belong with counsel qualified in professional responsibility rather than with an institute.

The reference material behind it covers ten areas, from choosing tools and finding the real productivity through verification, governance, billing and confidentiality, to what AI does to the business model and the training pyramid. It is written for managing partners and general counsel — competent lawyers with no obligation to care how a model works — and it is willing to say the uncomfortable thing, because a site that only said the comfortable things would not be worth reading.

The Institute also maintains a Library: the independent studies, the ethics opinions, the court decisions and the industry research, each read carefully and set against the others. Where an entry was published by an organisation that also sells legal AI, the entry says so. That is the whole editorial posture in one sentence.

plainly stated

What the Institute is and is not.

The Institute is

A knowledge institution. Plain-language references on tool evaluation, verification, governance, pricing, confidentiality, research, productivity, marketing and where this is all heading — written for the people who decide, not the people who build.
Independent of every vendor. The Institute takes no money from any company that builds or sells legal AI. It has no reseller relationship, no affiliate arrangement, and no product of its own to place.
Free at the point of use. The concierge, the references and the Library cost nothing and require no account.
Honest about what is unsettled. Jurisdictions have already diverged on fees and disclosure, the accuracy evidence moves between product releases, and nobody knows what happens to the associate pyramid. The Institute says so rather than manufacturing certainty.

The Institute is not

Not legal or ethics advice. Rules of professional conduct are adopted state by state and they diverge. The Institute describes what ABA Formal Opinion 512 addresses and what firms do in practice. It does not tell you what your jurisdiction requires, whether particular conduct is permissible, or whether something would be sanctionable — and anything involving an actual incident belongs with counsel qualified in professional responsibility, now.
Not a vendor ranking service. The Institute does not name a best tool, rank named products or recommend a vendor. Rankings age within months, cannot be validated against your matters, and could not be published credibly by anyone connected to people building in this space. You get the evaluation criteria instead, which is more useful anyway.
Not a technology implementer. The Institute does not sell you software, configure it, or take a position on your stack. If a decision comes down to which product, that is yours to make with the criteria in hand.
Not a party to your matter. Using this site creates no attorney-client relationship, and concierge conversations are neither privileged nor confidential. Do not put client names, privileged material or matter detail into it.

How the Institute sustains itself

The orientation is free and stays free. Where a firm wants sustained help, the Institute offers private advisory engagements: a readiness assessment, drafting or reviewing a firm AI policy, designing the verification step, working through what to bill for AI-assisted work, or helping leadership think through what this does to the firm’s economics. Those engagements are billed as fixed fees, agreed in writing before any work begins.

The Institute accepts no compensation of any kind from any company that builds, sells or resells legal AI — no sponsorship, no referral fee, no affiliate commission, no revenue share, no paid placement, and no free licences in exchange for coverage. No vendor has ever paid to appear on this site, because no vendor appears on this site.

We publish this because the Institute’s entire value is that it has nothing to sell you at the moment you are deciding what to buy. That claim is worth exactly as much as the evidence behind it, so here is the evidence: our fee does not change according to what you adopt, and does not change if you adopt nothing at all.

leadership

Leadership.

Announcement · August 2026

Russ Rosenzweig named Executive Director of the Institute

Russ Rosenzweig was named Executive Director of the AI for Legal Institute in August 2026. He was one of the pioneers of the expert witness industry and has decades of experience helping clients understand complex technical disputes and connecting them with the right experts and knowledge bases.

He founded the first expert witness search and referral firm in 1993 and led it for three decades, connecting thousands of attorneys, insurers and companies with specialized experts. His clients have included most of the largest law firms in the United States.

That history is the reason this institute is run from where it is. Three decades of serving law firms — through the arrival of electronic discovery, predictive coding and every technology the profession was told would change everything — is a long education in which changes actually reached practice and which did not. It also means the audience for this site is the audience he has spent a career listening to: the partner who has to decide, on incomplete information, with the firm’s name on the outcome.

B.A., Northwestern University · M.B.A., University of Chicago Booth School of Business

the network

The specialists behind the answers.

Adoption decisions rarely sit inside one discipline. A firm AI policy touches professional responsibility, information security, insurance and employment practice at once. Pricing AI-assisted work is a question for people who understand law firm economics, not technologists. Verification design is a knowledge-management problem. Getting any of it to survive a busy month is change management, which is its own field.

The Institute draws on a working network built over thirty years across those disciplines — professional-responsibility counsel, legal technologists, knowledge and innovation leaders, firm operations and finance people, and information security practitioners. Specialists are engaged matter by matter and remain independent. The Institute’s role is to find the right one quickly, and to assemble the group when a question crosses fields, which this one usually does.

An advisory council of firm leaders and professional-responsibility practitioners is being convened and will be named here.

standards

The rules this site holds itself to.

No vendor rankings, ever

Not a best tool, not a shortlist, not a quiet preference. This is written into the concierge’s operating rules and holds on every conversation, including when a visitor asks directly. You get criteria you can apply to your own matters instead.

The professional-responsibility line

The Institute describes what the guidance addresses and what firms do. It does not tell you what your jurisdiction requires or whether conduct is permissible. If something has already gone wrong, the concierge is instructed to say early and plainly that you need counsel in the relevant jurisdiction — first, not as a closing disclaimer.

Claims with a date on them

Accuracy rates change between product releases, adoption figures vary enormously by survey, and the sanctions record grows monthly. Every figure on this site carries the date it was measured, and none of them says “currently” or “the latest”.

Interests declared

Some of the best research in this field is published by organisations that also sell legal AI. The Institute cites that work and says so in the same breath, because leaving it out would be less useful than reading it with the interest in view.

AI, used openly

The Institute uses AI to research, draft and power the concierge, under human editorial direction. An institution telling law firms to be candid about their AI use would be in a poor position not to say so itself.

Corrections

Errors, once known, get fixed. If something on this site is wrong, tell the Institute and it will be reviewed and corrected.

Why this institute exists.