“He still prefers “enshittification” to “slop”, but the problem is both serious and ubiquitous no matter what you call it.”
Our reading, the sentence that was checked:
The problem of AI slop/enshittification is serious and ubiquitous.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Slopocalypse Now, paragraph 29.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-slopocalypse-now-6e1233-a2
We recorded 2 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 2025-08-04. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Observation
- Deadline, as the speaker wrote it
- None stated. This site does not add one.
- Deadline state when assessed
- No horizon
- Deadline state on 2026-10-29
- No horizon (the release’s reference date)
From the speaker
- the problem is both serious and ubiquitous
- no matter what you call it
- AI slop is everywhere
- journalism, law, music ... politics, education
Our interpretation
- Plain reading: by the post date (August 2025), low-quality or erroneous AI-generated content was widespread across many domains, not just one niche, and it was causing documented harm.
- 'Ubiquitous' is read as 'widespread across many domains', not literally present everywhere.
- 'Serious' is an evaluative word. It is read as 'causing documented real-world harm', such as sanctions on lawyers for fake AI citations or platforms flooded with AI content.
- No period is named. The claim describes an ongoing state, so the target period is January 2025 through the post date (2025-08-04).
Counts as supported if: Credible measurements or records from before the post showing large volumes of AI-generated content in several separate domains (for example the web, music, legal filings), plus documented harms.
Counts as contradicted if: Credible measurements showing AI-generated content was rare or confined to a few domains in mid-2025, or that documented harms were negligible.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Supported
The dated evidence matches what the claim committed to, under the criteria frozen before grading.
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 horizon
- Assessment recorded
- Policy version
- 2026-10-28.7
Rationale
The contract reads 'ubiquitous' as widespread across many domains and 'serious' as causing documented harm. It sets the period as January 2025 to the post date, 4 August 2025. Ahrefs reported that 74.2% of 900,000 new English-language web pages it analysed in April 2025 contained AI-generated content, based on its own detector (e1). Most of those pages were mixed, and 2.5% were rated pure AI. On 16 April 2025, Engadget reported Deezer's statement that about 20,000 fully AI-generated tracks were added each day, over 18% of daily uploads (e2). A post dated 25 May 2025 said Damien Charlotin's database listed 116 cases in 12 countries where a legal decision confirmed that a lawyer had presented AI-hallucinated content, 20 of them in May 2025 (e3). Together these show large volumes of AI content on the web and in music, plus recorded harm in legal filings, before the post. That meets the contract's support condition. The packet lists no counterevidence.
Limitations
- The result depends on the contract's reading of 'serious' and 'ubiquitous', which are evaluative or hyperbolic words.
- The Ahrefs figure counts any AI-generated content, mostly mixed with human writing. It does not measure low quality or slop.
- The web figure comes from Ahrefs' own detector, and the music figure is Deezer's own report as relayed by Engadget. The evidence does not show either was independently checked.
- The legal evidence is a secondary summary of one database and counts decisions, not the overall volume of AI use in filings.
- No cited measurement covers politics, journalism or education, which the post also names.
Evidence (3 items)
Supports the claim: Ahrefs,
We analyzed 900,000 newly created web pages in April 2025 and found that 74.2% of them contained AI-generated content.
Why it matters here: Shows AI-generated content was present in most new web pages in April 2025. This supports 'widespread'. Caveats: the measurement uses Ahrefs' own detector, and most pages were mixed human-and-AI rather than pure AI.
- Publisher
- Ahrefs
- Kind of source
- Independent measurement
- Describes events of
- Public from
Supports the claim: Engadget,
20,000 completely AI-generated tracks are added to its music service on a daily basis, making up "over 18% of all uploaded content" the platform receives every day.
Why it matters here: Shows that fully AI-generated tracks were a large and growing share of music uploads in a second domain. The same article says Spotify faces the same issue.
- Where
- Paragraph 1
- Publisher
- Engadget
- Kind of source
- News report
- Describes events of
- Public from
Supports the claim: Simon Willison's Weblog,
Damien's database has 116 cases from 12 different countries
Why it matters here: Records court decisions in 12 countries confirming that AI-hallucinated content was submitted in legal filings. This is documented harm in law, a third domain. The post also notes that 20 of the cases came in May 2025 alone.
- Where
- Paragraph 4
- Publisher
- Simon Willison's Weblog
- Kind of source
- Secondary summary
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Slopocalypse Now
He still prefers “enshittification” to “slop”, but the problem is both serious and ubiquitous no matter what you call it.
- Where
- paragraph 29
- Published
- 2025-08-04
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-slopocalypse-now-6e1233-a2: Current. Outcome: Supported. 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.