“Meanwhile, the companies are also not being held legally responsible for the damage that generative AI is causing to society.”
Our reading:
Generative AI companies are not being held legally responsible for societal damage caused by generative AI.
Assessment: Unresolved
What happened
Was it true for the period described?
Outcome: Unresolved
We checked, and the evidence so far doesn’t settle it.
Judged on a plain reading. They named no measure, so we used the reading a typical informed reader would take. It is spelled out under “How we judged it”.
Evidence up to . Checked .
Why
The record does not settle the fair contextual reading. Major copyright cases were still pending, which fits Marcus's concern about accountability. But France had penalised Google for breaches that included Bard-related conduct, and Italy had intervened against OpenAI. These are material qualifications. The Chinese copyright judgment also shows that provider liability existed, but an isolated case cannot defeat this generalisation about major firms. The evidence establishes neither broad accountability nor its general absence across the harms he describes.
Caveats
- The claim does not specify how much accountability would overturn its generalisation.
- The litigation survey and Chinese case account are secondary sources.
- Later judgments cannot change the historical assessment.
Evidence (4 items)
Context only: Chat GPT Is Eating the World,
OpenAI filed its motion to dismiss the first amended consolidated complaint of the book authors in In re ChatGPT Litigation (N.D. Cal.).
Why it matters: The publication-day survey documents major copyright proceedings still addressing motions to dismiss. This is consistent with delayed accountability in those cases, but cannot establish the absence of accountability across societal harms.
- Publisher
- Chat GPT Is Eating the World
- Kind of source
- Secondary summary
- About events on
- Published
Context only: Autorité de la concurrence,
Google has also proposed a series of corrective measures to address certain breaches identified by the Autorité de la concurrence.
Why it matters: The regulator imposed a €250 million penalty for several commitment breaches, including failures concerning Bard's use of publishers' content. This is material accountability involving a major provider, but the entire penalty cannot be attributed to generative AI and this decision alone does not establish broad accountability.
- Publisher
- Autorité de la concurrence
- Kind of source
- Regulatory filing
- About events on
- Published
Context only: Garante per la protezione dei dati personali,
OpenAI may submit its counterclaims concerning the alleged breaches within 30 days.
Why it matters: The regulator reported alleged privacy breaches following an earlier temporary processing ban, while reserving its final determination. This shows concrete regulatory intervention but does not establish completed, broad liability for the harms Marcus describes.
- Publisher
- Garante per la protezione dei dati personali
- Kind of source
- Regulatory filing
- About events on
- Published
Context only: Linklaters,
The Guangzhou Internet Court found a GenAI service provider liable for copyright infringement on 8 February 2024.
Why it matters: This later legal summary documents a pre-publication judgment concerning generated Ultraman images. It establishes an individual provider's liability, not broad accountability among the major firms discussed in the post.
- Where
- Opening paragraph
- Publisher
- Linklaters
- Kind of source
- Secondary summary
- About events on
- Published
Probably right?
One AI thinks this is probably right, and the other says it’s too early to tell. The evidence isn’t conclusive, so the verdict above stays unresolved. It counts only when you pick Speculative under “How sure?”.
- Anthropic’s AI: On publication day, the main copyright suits against OpenAI and Microsoft were still at the motion-to-dismiss stage, and Italy's privacy case was unresolved. The French €250M fine against Google, only partly about Bard, and a small Chinese judgment look like isolated exceptions.
How we judged it
- Kind of claim
- Observation
- Deadline
- None given. We don’t add one.
Where we had to interpret
- The ordinary contextual reading is that major generative AI companies were broadly escaping legal accountability for societal harms. This is not a universal assertion that no provider anywhere had ever lost a case.
- The surrounding discussion identifies Microsoft and OpenAI and includes harmful advice and harm to creators. Relevant accountability therefore includes binding compensation, penalties and corrective orders concerning development or deployment of generative AI.
- A lawsuit, investigation or allegation alone does not establish liability. Conversely, accountability need not take the form of damages paid to an individual.
- The observation concerns the legal position at publication. I use March 2024 through March 29 as the current reporting period, with earlier actions treated as historical context.
- The claim supplies no numerical threshold for broad accountability. Isolated decisions must be assessed for their relevance and reach rather than automatically deciding the claim.
Right if: A sufficiently broad record of major providers and relevant harms shows that meaningful legal consequences remained generally absent as of publication, with isolated exceptions insufficient to change that overall picture.
Wrong if: The record shows established, meaningful legal accountability across major providers or a substantial range of the harms described, making the general claim of escaping accountability untenable. One isolated judgment against a peripheral operator is insufficient.
We wrote this test after the outcome was known.
What would settle it
A sufficiently comprehensive historical review of judgments, binding settlements and regulatory remedies concerning major generative AI providers, showing their coverage of the described harms as of March 29, 2024.
We found 9 claims in this post. See them all.
Full record
Source and dates
Marcus on AI, The race between positive and negative applications of Generative AI is on – and not looking pretty, paragraph 14.
- Said
- We read the source
- live page, full text, public
- Evidence up to
- Checked
We read this wording on 2026-10-29. The post may have been edited after 2024-03-29.
Where it was said (1)
Said once in the posts we read.
: Marcus on AI, The race between positive and negative applications of Generative AI is on – and not looking pretty
Meanwhile, the companies are also not being held legally responsible for the damage that generative AI is causing to society.
- Where
- paragraph 14
- Published
- 2024-03-29
- We read it
- , live page
- Access
- full text, public
History
Versions
- Current: Unresolved. Evidence up to 2026-10-28.
What happened to this record
PublishedBy us
Published after both AI reviews and the final check passed.
Held backBy us
A reviewer said our test wasn’t fair to what was actually said.