Few professions have a sharper AI cautionary tale than law — courts have sanctioned lawyers for filing AI-fabricated case citations. Yet AI is also a powerful accelerant for research, review, and drafting. The difference is discipline: verification and confidentiality. Here’s how to adopt it right, and how dgm implements it. (dgm implements osFoundry, a separate company’s platform — we are not osFoundry; legal judgment and verification stay with your lawyers.)

What AI actually does for law firms

The honest framing: AI is a drafting, review, and research accelerant — it cuts the time lawyers spend on first-pass work — while legal judgment, client counsel, and accountability stay human. Used well it’s a genuine productivity gain; used carelessly it’s an ethics and malpractice risk.

High-value use cases

  • Legal research assistance — finding and synthesizing relevant material (lawyer-verified).
  • Document review and e-discovery — surfacing relevant documents at scale.
  • Contract drafting and review — flagging clauses and deviations, drafting first passes.
  • Client-intake support — handling routine intake and questions.

The ethics and verification reality

This is where law-firm AI lives or dies:

  • ABA Formal Opinion 512. The ABA’s first formal guidance on generative AI (July 2024) frames the duties: competence (a reasonable understanding of the tools), confidentiality (Rule 1.6 — generally obtain client informed consent before inputting client confidences; boilerplate consent isn’t adequate), communication, candor to the tribunal, supervision, and reasonable fees. State bars have issued guidance too.
  • Hallucinated citations are real and sanctioned. In Mata v. Avianca (2023), a federal court sanctioned attorneys who filed a brief citing fictitious cases fabricated by ChatGPT — and courts have addressed many similar cases since. The rule is absolute: every AI-provided citation must be verified by a lawyer before filing.
  • Client confidentiality with AI vendors means controlling where client information goes and ensuring vendors don’t train on it.

dgm builds confidentiality controls and verification workflows into the implementation; legal judgment and verification stay with your lawyers.

How to start

Start with document review or research — high-volume work — with mandatory citation verification and client-confidentiality controls in place. Never file AI output unchecked. Prove the time saved, then expand. dgm’s assessment finds the right starting point. (Note: specialized legal-vertical AI tools also exist — see osFoundry vs Harvey — and dgm can help you weigh general vs specialized.)

How dgm helps

dgm implements osFoundry and other AI for US law firms — within client-confidentiality controls and with citation-verification workflows built in, and training your team. Pricing is fixed and public: a $399 assessment and $3,999/month implementation, with no per-seat fees. If you’d rather explore the platform first, go straight to osFoundry; if you want law-firm AI done right and ethically, that’s where dgm comes in.