“Be on your guard; it’s going to happen a lot.”
Our reading, the sentence that was checked:
Covert opt-out scraping of user data for AI training will happen a lot.
Assessment: Unresolved No deadline stated on 2026-10-29 Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, PSA: Microsoft may be training on your private data without your knowledge, paragraph 8.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-psa-microsoft-may-be-training-on-0d52b0-a2
We recorded 6 claims from this post. See every claim from this post.
The wording above is the version this site observed on 2026-10-28. The source may have been edited after 2024-11-26. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Forecast
- Deadline, as the speaker wrote it
- None stated. This site does not add one.
- Deadline state when assessed
- No deadline stated
- Deadline state on 2026-10-29
- No deadline stated (the release’s reference date)
From the speaker
- it’s going to happen a lot
- the same sly opt-out playbook
- hiding what is going in obscure settings and Terms of Service
Our interpretation
- It refers to the preceding prediction of concealed default inclusion of user data in AI training.
- A lot means a recurrent, widespread practice across companies and services, rather than an isolated case.
- No numerical count, market share, frequency, or deadline is supplied. I do not invent one.
- Be on your guard is advice and is not itself graded.
Counts as supported if: Dated evidence establishes a broad, repeated post-publication pattern of companies using default inclusion for AI training while leaving material disclosure and opt-outs obscure.
Counts as contradicted if: A decisive development rules out the predicted widespread recurrence. A limited number of examples or a period without examples cannot refute an undated forecast.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Did the forecast come true?
Outcome: Unresolved
The record was checked, and the evidence available by the cutoff does not settle it.
Scored on a plain reading. The speaker named no measure for this claim, so it was judged on the reading an informed reader would take, stated below under our interpretation. Both reviews had to find that reading fair.
- Evidence as of
- Deadline state when assessed
- No deadline stated
- Assessment recorded
- Policy version
- 2026-10-28.7
Rationale
Open Terms Archive documented a January 16, 2026 Starlink policy change permitting AI training and requiring users to opt out of third-party training [e1]. Anthropic's August 28, 2025 announcement said existing users would see a training choice in a pop-up window [e2]. These records establish neither a broad, repeated pattern of default training use with obscure disclosure nor a decisive development ruling that pattern out [e1, e2].
Limitations
- Starlink's policy permissions do not establish actual training use or the visibility of notices to affected users.
- Anthropic's announcement describes its planned user-choice process, not independently verified delivery or user understanding.
- The evidence does not establish prevalence across companies and services.
- The forecast supplies no deadline or numerical threshold for 'a lot.'
Evidence (2 items)
Context only: Open Terms Archive,
January 16, 2026
Why it matters here: The policy monitor documents training permissions and an opt-out requirement for third-party training. One observed policy change cannot establish widespread covert training or actual use of the data.
- Publisher
- Open Terms Archive
- Kind of source
- Independent measurement
- Describes events of
- Public from
Context only: Anthropic,
a pop-up window
Why it matters here: Anthropic describes presenting the training choice directly to existing users. This challenges counting its policy change as evidence of a concealed practice, but cannot refute a forecast about other companies.
- Publisher
- Anthropic
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
What would settle it
A dated cross-company investigation or substantial collection of verified records establishing a broad, repeated post-publication pattern of default inclusion in AI training with obscure notices and opt-outs; alternatively, a decisive development ruling out the predicted recurrence.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, PSA: Microsoft may be training on your private data without your knowledge
Be on your guard; it’s going to happen a lot.
- Where
- paragraph 8
- Published
- 2024-11-26
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-psa-microsoft-may-be-training-on-0d52b0-a2: Current. Outcome: Unresolved. Evidence as of 2026-10-28. Public date not recorded.
Recorded events
Assessment publishedBy this site
First assessment published after both reviews and the final-text check passed.
About assessment marcus-psa-microsoft-may-be-training-on-0d52b0-a2.