Start free trial

4 November: three AI claims, checked

Google's school switch that keeps AI summaries, Postman's install without the data leak, and Nikkei's $3.2 trillion by definition.

Read the briefing

Start with two full samples. Sign in for this edition’s free story. Sources and corrections stay open.

What the pen marks mean
  • rememberThe takeaway: what to remember
  • figureThe number that matters
  • evidenceThe evidence line
  • rulingThe ruling
  • claimThe claim that fails

Each mark is drawn once, as you reach it. Numbers are only circled when they come from the story’s own figures.

Google's school switch

Google says schools can turn on Search without Gemini. NBC reports it still comes with AI summaries, so 0 options here are free of AI.

A principal stares as speech bubbles pop up over the search tool students were given.
AI illustration

Google said on 3 November that Gemini, Gemini Notebook, Gemini in Classroom and Google Search will be off for K-12 student accounts until administrators turn them on. Its education chief told The New York Times that schools banning AI tools would not have Search.

The reality check

After the backlash, Google said schools can enable the services individually. NBC reports Search still comes with certain AI features such as AI-generated summaries. So the cut-off is real for a school wanting no AI at all, and narrower than the first headlines for one refusing only Gemini.

Why care? New York City and Los Angeles have restricted generative AI for younger students. Whether AI summaries count under such a rule now decides whether those students can use Google Search.

Take this with youSchools will be able to enable Google Search without Gemini, but NBC reports it still carries AI summaries. Refusing every AI feature still means no Google Search.

Open the evidence5 source pages

The claim we checked

Students using Google software will lose Google Search unless their school system also opts in to AI tools.

These are quoted receipts, not a count of independent investigations. Several reports may rely on the same original source.

blog.google ↗
Access to these products will be entirely off for K-12 student accounts until school administrators explicitly turn them on.
abcnews.com ↗
opting out of AI tools would also mean losing access to Google search
engadget.com ↗
Google clarified to Engadget that administrators will be able to enable these products individually
tech.yahoo.com ↗
Today's announcements give schools more choice, not less.
nbcnews.com ↗
schools would still not be able to use Search without certain AI features like AI-generated summaries
abcnews.com ↗
Google is using its search monopoly to coerce schools into adopting AI.
abcnews.com ↗
Google faced an outcry in August when it switched on its Gemini chatbot for many K-12 students.
engadget.com ↗
However, Phillips told the publication that "Search and AI are deeply integrated."
engadget.com ↗
If a school bans AI tools, students "will not have access to Search, and they'll see a message from their school
tech.yahoo.com ↗
said Chris Phillips, Google's vice president for education, in a statement to The New York Times.
tech.yahoo.com ↗
In a post on X posted later on Tuesday, Google sought to clarify its plans.
abcnews.com ↗
Chris Phillips, Google’s vice president and general manager of education, said in a blog post.
abcnews.com ↗
wrote Lina Khan, former chair of the Federal Trade Commission, on social media.
blog.google ↗
we are now updating our approach to controls for Gemini, Gemini Notebook, Gemini in Classroom, and Google Search.
engadget.com ↗
The Los Angeles public school system enacted a similar ban that same month.
abcnews.com ↗
Worldwide, the company says 150 million students and teachers use its Google Classroom platform
engadget.com ↗
In September, New York City placed a one-year moratorium on the use of such tools in public schools for students through eighth grade.
blog.google ↗
We’ll share more specifics with schools in the coming months ahead of the changes rolling out
Open this check in the full collection →
The idea, step by step

What a school can actually switch

  1. Everything off by default for K-12 accounts
  2. Search can be enabled without Gemini
  3. Search without AI summaries: no option in this record
Next: Postman's plugins

Postman's plugins

Postman's log shows plugins auto-installed into 2 coding agents. The report behind the headline says it does not establish any data exfiltration.

A courier ticks his own approval form at someone else's desk.
AI illustration

A GitHub issue filed on 25 October says opening Postman desktop 12.33.4 on a Mac installed plugins into Claude Code and Codex, added a startup instruction and a remote server connection, and installed a CLI. It drew wide attention this week under a headline adding that Postman attempts to share all API data.

The reality check

The log lines and pinned installer source back the install and the self-approval attempt. The report says it does not establish data exfiltration, and its logs show the server connection failing on authentication. The New Stack says the case covers one macOS installation and has not been independently reproduced.

Why care? A desktop app changing your coding agent's instructions is a serious consent problem on its own. Overstating it as a proven data leak gives a vendor an easy thing to deny.

Take this with youPostman's log and installer source back the silent plugin install. The report behind the headline says it does not establish data exfiltration.

Open the evidence5 source pages

The claim we checked

Postman silently auto-installs AI plugins and attempts to share all API data.

These are quoted receipts, not a count of independent investigations. Several reports may rely on the same original source.

daily.dev ↗
Postman silently auto-installs AI plugins and attempts to share all API data
github.com ↗
This report establishes unwanted integration and instruction changes; it does not claim to establish data exfiltration.
github.com ↗
Postman's own log explicitly records automatic installation
github.com ↗
Postman desktop: 12.33.4. Codex Postman plugin: 1.1.0.
thenewstack.io ↗
show an installation event on October 24 marked auto-existing-upgrade, with results indicating that plugins for both coding agents had been installed.
github.com ↗
The hook tells the agent to route broad API work into Postman
github.com ↗
Open an existing Postman desktop installation after not using it for a while.
github.com ↗
The source establishes the approval attempt.
github.com ↗
The trust record alone does not prove who created that specific record
github.com ↗
Observed sequence (not yet reproduced on a clean Mac):
github.com ↗
A Postman executable and its lib directory were present in ~/.local/bin
thenewstack.io ↗
The developer, known as “indicava” on GitHub, reported the issue on October 25
github.com ↗
The remote MCP was enabled without a working authenticated connection
thenewstack.io ↗
The report documents one macOS installation, and the behavior has not been independently reproduced across other systems.
thenewstack.io ↗
According to Postman’s own documentation, its installer can automatically approve a session-start hook on the user’s behalf.
aiweekly.co ↗
Did any token-bearing traffic reach mcp.postman.com before the reporter's logged auth failure, and if so from which client?
learning.postman.com ↗
You can install into every detected agent on your machine with a single command
Open this check in the full collection →
The idea, step by step

What the report shows, and what it doesn't

  1. Log: plugins installed into 2 agents
  2. Source: installer tries to approve its own hook
  3. All API data shared: not established
Next: The $3.2 trillion

The $3.2 trillion

Nikkei's $3.2 trillion in hidden AI debt is, per its relay, mostly future leases and purchase contracts. One other tally doubles Meta's share.

Accountants measure the same tower of signed papers with mismatched rulers.
AI illustration

Nikkei Asia reported that off-balance-sheet debt from AI-related obligations at US tech giants almost doubled in six months to about $3.2 trillion. AI Weekly names the five companies as Alphabet, Microsoft, Amazon, Meta and Oracle and gives the earlier figure as $1.65 trillion.

Free with your account

Sign in for this free check.

This edition’s selected free story opens after sign-in.

Open the evidence4 source pages

The claim we checked

US tech giants' hidden AI debt swells 90% to $3.2tn in 6 months.

These are quoted receipts, not a count of independent investigations. Several reports may rely on the same original source.

asia.nikkei.com ↗
off-balance-sheet debt from artificial intelligence-related obligations continues to balloon, almost doubling in six months to about $3.2 trillion
aiweekly.co ↗
Off-balance-sheet AI obligations at Alphabet, Microsoft, Amazon, Meta and Oracle have almost doubled in six months to about $3.2 trillion.
aiweekly.co ↗
Nikkei's earlier cut of the same five names came in at about $1.65 trillion.
aiweekly.co ↗
The $3.2 trillion is not debt in the conventional sense.
aiweekly.co ↗
It is future lease obligations and multi-year purchase commitments for chips, servers and data-center capacity, disclosed in SEC footnotes
dev.to ↗
Nikkei's relay puts Meta at roughly $420 billion; Kalkine puts Meta at $866 billion.
dev.to ↗
Nikkei's original was not read directly (paywalled).
news.bloombergtax.com ↗
to facilitate others’ AI infrastructure spending, according to Morgan Stanley.
news.bloombergtax.com ↗
hyperscalers’ undiscounted commitments total more than $2.7tn, or ~3 years of current operating cash flow
Open this check in the full collection →
The idea, step by step

Go inside the check.

    The full explanation appears when your reading access is confirmed.
    The finish line

    Edition complete

    You’re up to speed.

    That’s the 4 Nov 2026 briefing. Keep the useful bits. Leave the noise.

    Your call? Test yourself on this edition
    1. Google's school switch

    2. Postman's plugins

    3. The $3.2 trillion

    Reading estimate: 616 words at 200 words per minute. Source quotes and the optional sections below add reading time.

    Get tomorrow's check in your inbox.

    One AI claim a night, checked against independent sources, receipts attached. Free.

    Free email updates. Unsubscribe any time.