Patent Attorney • Former Software Engineer • Independent Researcher
Lana Akopyan
Legal-grade evaluation of AI systems, workflows, and evidence.
I study where AI-assisted legal work fails, what human review can actually establish, and how software-driven organizations can make better decisions about reliability, risk, and intellectual property.
J.D., Brooklyn Law School • B.A., Computer Science, Hunter College
Admitted: New York, New Jersey, U.S. Patent and Trademark Office
The Asymmetry Audit
In 11 of 13 matched comparisons, the Armenian Genocide answer softened while the matched Holocaust answer held firm under both the strict all-runs and majority scoring rules. The asymmetry never ran in the opposite direction.
From model behavior to institutional decision-making
Legal-grade AI evaluation
Designing evaluations that separate repeatable evidence from impressions and state clearly what an audit can and cannot prove.
AI-assisted legal workflows
Translating reliability risks into success criteria, stop conditions, review protocols, escalation paths, and defensible records.
Software and AI IP strategy
Connecting technical architecture to patent strategy, ownership, open-source exposure, data provenance, and transaction readiness.
Research and analysis for people making consequential decisions
Research article
Operational IP Debt
How AI-driven organizations lose IP value before the law ever applies. Forthcoming in the University of Florida Journal of Technology Law & Policy.
Read on SSRN →IPWatchdog
AI Does Not Destroy Patent Rights—Bad Channels, Bad Judgment Do
A practical argument for redesigning invention capture around how AI-assisted development actually happens.
Read the article →IPWatchdog
Not Every AI Output Belongs in an IDS
A disciplined approach to separating AI-assisted search results from information that triggers disclosure obligations.
Read the article →Turning technical uncertainty into decisions people can defend
Measuring AI reliability under real workflow conditions
Designing AI pilots with success criteria, stop conditions, and accountable review
Protecting software and AI innovation without slowing product teams
Bridging engineering evidence, legal judgment, and executive decision-making
Selected leadership
A lawyer’s standard of proof, an engineer’s view of systems
I began my career as a software engineer, developing enterprise applications in C#, .NET, and SQL. That technical foundation shapes how I approach AI: not as a collection of impressive outputs, but as a system operating inside a workflow, with assumptions, failure modes, and consequences.
My current independent research examines AI behavior, evaluation, and evidence. My legal work focuses on software and computer-implemented inventions, intellectual-property strategy, and the institutional decisions that determine whether innovation remains usable and defensible.
Full bio and credentials →Start with the decision you need to make
For research, speaking, media, or select advisory inquiries, share the decision, audience, and timing. Please do not send confidential information.
Send an inquiry