Insights · IP exposure · 30 Apr 2026

Norton Rose Fulbright, 2025:
47% of those expecting more IP exposure point to AI tools.

The annual litigation-trends survey that in-house legal teams at IP-heavy companies watch for. What the 2025 edition said about AI tools, trade secrets, and where the next IP dispute comes from.

47%

of respondents expecting more IP exposure point to AI tools as a major factor

 44%

of those who saw IP exposure grow trace it to trade secrets

 NORTON ROSE FULBRIGHT · 2025 ANNUAL LITIGATION TRENDS SURVEY
Two statistics from the Norton Rose Fulbright 2025 Annual Litigation Trends Survey: 47 percent of respondents expecting more IP exposure point to AI tools as a major factor; trade secrets accounted for 44 percent of those who experienced IP exposure growth.

Norton Rose Fulbright’s Annual Litigation Trends Survey is a long-running benchmark in corporate legal departments. The 2025 edition surveyed in-house counsel across the United States and Canada. On IP and AI, the survey is specific:

“Use of third-party software or components (such as free AI tools) is a major factor driving dispute exposure, selected by almost half (47%) of respondents who anticipate an increase in IP exposure in the year ahead. … Trade secrets accounted for 44% of those who experienced IP exposure growth.”
Norton Rose Fulbright, 2025 Annual Litigation Trends Survey

The mechanism is the one the Samsung incident made famous in April 2023: three separate leaks of source code, equipment yield data, and internal meeting notes into ChatGPT within one month. Samsung banned the tool company-wide by May 1, 2023. The litigation-trends data says the same pattern is now showing up in dispute pipelines across both Canadian and US in-house legal departments.

The supporting context comes from ShareGate’s March 2026 survey of 851 IT and security leaders: nearly 1 in 3 organizations across North America and Europe report that AI tools have already surfaced sensitive data they shouldn’t have accessed. The same loss pattern runs through the Cassels “Trade Secrets Year in Review” and the Bennett Jones commentary on the U.S. v. Heppner ruling (2025), the case that held AI communications with Claude lack confidentiality protection.

Read the source

Why this matters

If you are a General Counsel, CLO, or CTO at an IP-critical organization (technology, pharma, life sciences, engineering, mining, management consulting), the survey names the risk directly. The asset your company was built on, whether that is source code, formulations, clinical protocols, or proprietary methods, can leave through an unmonitored AI prompt, and you cannot tell the acquirer, the investor, or the auditor what left. For the AI traffic you route through it, Mandate produces the structured record of which tool was used, the content category it touched, and the policy decision that applied.

More from Insights

Preparing for due diligence?
The record starts the day the proxy goes live.

A 30-day pilot puts a structured, hash-chained record of the AI traffic you route through Mandate on the calendar before the next investor or acquirer asks.

Read about the pilot Get in touch

contact@mandateco.ca  ·  1-905-630-1908