What Lawyers, Researchers, and Authors Have in Common Now

A lawyer in New York. A researcher in London. An author in Melbourne. Three people in three professions on three continents, and in 2026, they all face the same question:

"Can you prove your thinking shaped this work?"

The lawyer

After Mata v. Avianca, the legal profession woke up. A lawyer submitted a brief with AI-hallucinated case citations. He didn't verify them. The court sanctioned him. The ABA responded with Formal Opinion 512: lawyers must disclose AI use and ensure the accuracy of AI-assisted work.

But the opinion goes further than most people realize. It doesn't just say "disclose." It implies that lawyers should be able to demonstrate the process by which they verified AI outputs. That's not a checkbox. That's an audit trail.

The researcher

ICMJE, the body that sets authorship guidelines for medical journals, now requires that researchers document AI's role in their work. NeurIPS, the leading AI conference, demands similar disclosure for paper submissions.

But here's the problem: a researcher who used AI to help structure their literature review and a researcher who used AI to generate their entire methodology section would write the same disclosure statement. The documentation that matters, the record of what the AI contributed and what the human decided, doesn't exist in current tools.

The author

Publishers are implementing AI policies at speed. Some ban AI outright. Others allow it with disclosure. But the authors caught in between, the ones who genuinely used AI as a thinking tool without letting it write for them, have no way to prove it.

"I only used it for brainstorming" is not evidence. It's a claim. And in an era where AI can produce entire novels, claims without evidence carry less weight every year.

The shared gap

All three professions need the same thing: a verifiable record of the human's cognitive contribution to AI-assisted work. Not a disclosure form. Not a timestamp. A structured, tamper-evident audit trail that shows:

• What was asked of the AI
• What the AI returned
• What the human did with it
• What was adopted, modified, or rejected

This infrastructure doesn't exist in ChatGPT, Grammarly, Copilot, or any other mainstream AI tool. Which means that across law, research, and publishing, professionals are being asked to meet accountability standards with no tools to help them do it.

Why this matters

The convergence is not coincidental. As AI becomes capable enough to produce professional-quality work, every field that values human expertise faces the same structural problem: how do you prove the human mattered?

The answer is the same across all three: you build the infrastructure to make it visible.