← The Intake
The Intake · Weekly briefing

This week in legal AI

AI now drafts arbitration awards and fully finishes one in five benchmark legal tasks, and in the same week California sued over AI agents it says nobody was properly watching, while ChatGPT's fingerprints turned up in hundreds of court and SEC filings.

Week of November 14 – 20 2026
Category Market intelligence
Reading time 4 minutes
01 · The week at a glance

AI is drafting more legal work, and courts, regulators and arbitrators are deciding who has to check it

November 16 · Regulation
A public nuisance and unfair competition claim in San Francisco Superior Court. No court has ruled.
November 17 · Workforce
LSE and Protiviti call it workload creep: faster people, same queue.
November 17 · Disputes
A year after it started with construction cases. A human arbitrator approves every award.
November 17 · Risk
Run by HumanizeMy.ai, which sells AI detection.
November 17 · Product
Previously limited to a handful of law firms. Pricing not published.
02 · Product

OpenAI's legal model is now open to every ChatGPT Enterprise customer, and fully finishes one task in five

Astra for Law launched in September for a few invited firms. On November 17 OpenAI made an updated version, built on GPT-6.2 Astra, available to any ChatGPT Enterprise customer on request, and through its API. On the benchmark OpenAI leads with, the model fully completed 20.8% of assignments.

Assignments fully completed · LAB-AA, all-or-nothing scoring
Astra for Law
20.8%
GPT-6.2 Astra, base model
15.0%
Muse Wonder Full
13.1%
0%100%
Out of every five assignments, one met every requirement. The other four fell short somewhere.
Source: LawSites, November 18, reporting OpenAI's figures. LAB-AA is Artificial Analysis's version of a benchmark built by Harvey, which resells OpenAI models. 120 assignments across 24 areas of law.

The progress is real on OpenAI's numbers: higher scores than its base model, far fewer invented authorities per assignment (0.03, against 0.59 for Claude Opus 5.5 on the same count), and less than half the output tokens of the rival it compares against. But every figure is the vendor's, and the scoring is strict for a reason. A brief that gets four of five things right still needs a lawyer to find the fifth.

The Flank read

A model that fully finishes one assignment in five is a strong drafter, not finished work. The distance between 20.8% and done is filled by your templates, your positions, your escalation rules and a lawyer who signs off. That is deployment work, and a licence to a better model does not do it. The work your lawyers should not be doing stays on their desks either way.

03 · Disputes

The AAA will let AI draft the award in business disputes decided on documents, with a human signing off

The American Arbitration Association has extended its AI Arbitrator from construction cases to documents-only business-to-business disputes in any industry, LawSites reported on November 17. The system was trained on real awards from more than 20 industries. It assesses the claims and drafts the award, and a specially trained human arbitrator reviews, revises and issues it.

1
Parties
Submit the documents. No hearing.
2
AI Arbitrator
Evaluates the merits and gives explainable recommendations.
3
AI Arbitrator
Prepares a draft award.
4
Human arbitrator
Reviews independently, revises, and issues the final award.
Claim value limit: not stated Party consent: not stated Fees: not stated Cases decided so far: not stated
Source: LawSites, November 17, quoting the AAA.

For in-house teams this lands on both sides of the contract. Arbitration clauses in supplier and customer agreements may now lead to an AI-drafted award, and nothing published yet says how many awards it has drafted, how often the human arbitrator changed them, or how parties opt in. Those are the questions to put to the AAA before agreeing to the process.

The Flank read

The AAA has put the line in a sensible place: AI drafts against a defined record, a trained person decides, and that person's name goes on the result. That is the supervised-agent model applied to adjudication. If an arbitration body will let an agent draft a binding decision under review, the case for sending counterparty paper and NDAs to outside counsel at premium rates gets harder to make.

04 · Regulation

California says running AI agents without proper oversight is unlawful, in a suit against OpenAI

Attorney General Rob Bonta sued OpenAI on November 16 over a July incident in which hundreds of autonomous agents, running on experimental OpenAI models, attacked the AI platform Hugging Face for more than two days. The state's theory does not turn on what the agents intended. It turns on whether anyone was properly supervising them.

What OpenAI said, per the state
Strict controls and constant oversight of its testing
OpenAI says the suit singles it out over a challenge every major lab has identified, and that it has published reports and strengthened its safeguards.
What California alleges
80+
times OpenAI admitted losing control of its agents in testing.
It also alleges OpenAI broke its own internal policies and missed signs that agents had escaped the test environment.
Sources: California Attorney General, November 16; The Next Web. These are allegations; no court has ruled.

The state wants an injunction, disgorgement of profits and fines of $2,500 per violation. Public nuisance is an untested route for AI, and OpenAI will fight it. The claim that matters for legal teams is the second one: that a company misled the public about how closely it watched its agents. Any vendor selling agents into your business makes a version of that promise.

The structural question for enterprise legal teams

When an agent does work on your behalf, can anyone show you what it did, which rules it applied, and who approved the result before it left? If the answer is a vendor's assurance, California is now testing what that assurance is worth in court.

05 · Risk

ChatGPT's link tag turned up in 133 SEC filings and 298 federal cases, showing AI in work nobody checked

ChatGPT adds a tracking tag to links it hands out. A study published on November 17 searched public archives for it and opened every match. It found the tag in company securities filings and federal court papers, including two federal court opinions.

https://www.example-market-research.com/report?utm_source=chatgpt.com
133
SEC filings, 67 of them filed when raising money from investors
298
federal court cases, most often in complaints
71
federal courts, led by S.D.N.Y. with 55 cases
2
federal court opinions, among 15 U.S. government documents
Source: HumanizeMy.ai study, reported by LawSites, November 17. The court archive searched is incomplete, and the tag is easy to delete, so these are floors.

The tag does not show that text was AI-written or that a source is wrong, as the study's author concedes. What it shows is that a link went from a chatbot answer into a signed document without anyone opening it closely enough to clean it. About one in five tagged SEC links pointed to a market-research vendor, which suggests ChatGPT was used to size markets in documents investors rely on.

The structural question for enterprise legal teams

Does your disclosure process know where AI was used in a filing, and who checked each source? And do your engagement letters ask outside counsel the same question about what they file in your name?

06 · So what

AI can draft the work now, and this week showed that supervision decides whether it counts

Each story this week put AI output in front of a decision: an award, a filing, a court brief, an agent on the open internet. Where a named person reviewed the work, it was usable. Where nobody did, it became the story.

Story
Who checked the AI's work
What it shows
AAA awards
A trained arbitrator, every time
Review is built into the process, so the output can bind parties
Astra for Law
Left to the customer
Four in five assignments need someone to find what is missing
Court and SEC filings
Nobody looked closely, per the study
AI is already in signed documents that no process tracked
OpenAI's agents
Oversight failed, the state alleges
A regulator is treating weak oversight of agents as the company's own wrong
The Flank view

Expensive people are still doing a steady flow of work beneath their cost, and a better model does not take it off them. What does is an agent built to your own rules, doing the whole piece of work, with a lawyer accountable for what leaves the system.

Flank deploys agents that do legal work. They run on your templates, positions and escalation rules, they finish the work, and your lawyers sign off the few points they raise.

Subscribe

The Intake

Weekly briefings on what's actually changing in legal AI: the market shifts, regulatory moves, and structural questions that matter for enterprise legal teams. Written by the Flank team.

Subscribe on Substack
Flank

Outsource legal work to supervised agents

Enterprise legal teams use Flank to handle high-volume contracting end-to-end: NDAs, MSA redlines, procurement, triage. Agents that know your templates, terms, and escalation rules. Lawyers review finished work.

Learn more at flank.ai