“While I’d love to believe that Apple has only used licensed content to train its models, that clearly isn’t the case, and if their version of “publicly available” includes things from Google, or DeviantArt, or social networks, Apple is willingly participating in the AI boom’s continual pillaging of the internet.”
Our reading:
Apple has not trained its models only on licensed content.
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
Apple's June 10, 2024 disclosure describes training on licensed data and publicly available data collected by AppleBot. [e1] It does not identify an included work used without an applicable license or provide a complete licensing inventory. [e1] The disclosure therefore establishes web collection but does not settle whether the relevant training completed by June 17, 2024 used any unlicensed content. [e1]
Caveats
- Public accessibility and crawler collection do not establish licensing status.
- The disclosure chiefly covers two language models and does not provide a complete account of image-model training.
- The evidence does not establish whether any training use infringed copyright.
Evidence (1 item)
Context only: Apple Machine Learning Research,
opt out
Why it matters: The surrounding paragraph distinguishes licensed datasets from publicly accessible material collected by AppleBot. That establishes web collection, but does not identify an included work's licensing status or establish that every crawled work lacked a license.
- Publisher
- Apple Machine Learning Research
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- About events on
- Published
How we judged it
- Kind of claim
- Observation
- Deadline
- None given. We don’t add one.
Where we had to interpret
- Assess the factual assertion about the training of the Apple Intelligence models discussed in the post.
- Licensed includes applicable open licenses as well as individually negotiated licenses.
- Public accessibility and collection by a crawler do not, by themselves, establish whether content was licensed.
- The conditional speculation about particular websites and the moral characterization of pillaging are not separate factual commitments in the normalized claim.
- The target is training completed by the post date; no training start date is supplied.
Right if: A reliable disclosure or documented training record establishes inclusion of content used without an applicable license.
Wrong if: A sufficiently complete training-data and rights record establishes that all relevant training content was licensed.
We wrote this test after the outcome was known.
What would settle it
Reliable records showing that content used in the relevant models' training by June 17, 2024 lacked an applicable license, or a sufficiently complete training-data and rights inventory establishing that all relevant content was licensed.
We found 19 claims in this post. See them all.
Full record
Source and dates
Where's Your Ed At, Let Tim Cook, paragraph 19.
- Said
- We read the source
- live page, full text, public
- Evidence up to
- Checked
We read this wording on 2026-10-28. The post may have been edited after 2024-06-17.
Where it was said (1)
Said once in the posts we read.
: Where's Your Ed At, Let Tim Cook
While I’d love to believe that Apple has only used licensed content to train its models, that clearly isn’t the case, and if their version of “publicly available” includes things from Google, or DeviantArt, or social networks, Apple is willingly participating in the AI boom’s continual pillaging of the internet.
- Where
- paragraph 19
- Published
- 2024-06-17
- 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
The final check failed: a cited source didn’t say what the write-up said, and the write-up said more than the evidence showed, and a condition they stated was left out.