> We believe in training our models using diverse and high-quality data. This includes
data that we’ve licensed from publishers, curated from publicly available or open-
sourced datasets, and publicly available information crawled by our web-crawler,
Applebot.
> We do not use our users’ private personal data or user interactions when training
our foundation models. Additionally, we take steps to apply filters to remove certain
categories of personally identifiable information and to exclude profanity and unsafe
material.
> Further, we continue to follow best practices for ethical web crawling, including
following widely-adopted robots.txt protocols to allow web publishers to opt out
of their content being used to train Apple’s generative foundation models. Web
publishers have fine-grained controls over which pages Applebot can see and how
they are used while still appearing in search results within Siri and Spotlight.
When Apple inevitably partners with OpenAI or Anthropic, which by their definition isnt doing "ethical crawling", I wonder how I should be reading that.
I mean they also buy from companies with less ethical supply chain practices than their own. I don’t know that I need to feel anything about that beyond recognizing there’s a big difference between exercising good practices and refusing to deal with anyone who does less.
You shouldn't believe Big Tech on their PR statements.
They are decades behind in AI. I have been following AI research for a long time. You can find best papers published by Microsoft, Google, Facebook in past 15 years but not Apple. I don't know why but they didn't care about AI at all.
Apple used to be at the edge of AI. They shipped Siri before "AI assistant" went mainstream, they were one of the first to ship an actual NPU in consumer hardware and put neural networks into features people use. They were spearheading computational photography. They didn't publish research, they're fucking Apple, but they did do the work.
And then they just... gave up?
I don't know what happened to them. When AI breakthrough happened, I expected them to put up a fight. They never did.
>I don't know what happened to them. When AI breakthrough happened, I expected them to put up a fight. They never did.
Apple always had the luxury of time. They work heavily on integrating deeply into their ecosystems without worrying about the pace of the latest development. eg. Widgets were a 2023 feature for iOS. They do it late, but do it well.
The development in the LLM space was and is too fast for Apple to compete in. They usually pave their own path and stay in their lane as a leader.
The impact on Apple's brand image will be tarnished if Google, Meta, OpenAI, MS all leapfrog Apple's models every 2-3 months. That's just not what the Apple brand is associated with.
One problem with Apple's approach here is that they were scraping the web for training data long before they published the details of their activities and told people how to exclude them using robots.txt
That seems like a potentially very useful addition to the robots.txt "standard": Crawler categories.
Wanting to disallow LLM training (or optionally only that of closed-weight models), but encouraging search indexing or even LLM retrieval in response to user queries, seems popular enough.
If you're using a specific user agent, then you're saying "I want this specific user agent to follow this rule, and not any others." Don't be surprised when a new bot does what you say! If you don't want any bots reading something, use a wildcard.
Yes, but given the lack of generic "robot types" (e.g. "allow algorithmic search crawlers, allow archival, deny LLM training crawlers"), neither opt-in nor opt-out seems like a particularly great option in an age where new crawlers are appearing rapidly (and often, such as here, are announced only after the fact).
Sure, but I still think it's OK to look at Apple with a raised eyebrow when they say "and our previously secret training data crawler obeys robots.txt so you can always opt out!"
I've been online since before the web existed, and this is the first time I've ever seen this idea of some implicit obligation to give people advance notice before you deploy a crawler. Looks to me like people are making up new rules on the fly because they don't like Apple and/or LLMs.
It's not controversial, it's just not how the ecosystem works. There has never been an expectation that someone make a notification about impending crawling.
It might be nice if there were categories that well-behaved bots could follow, as noted above, but even then the problem exists for bots doing new things that don't fall into existing categories.
My complaint here isn't what they did. It's that they explain it as "here's how to opt out" when the information was too late to allow people to opt out.
> Using our web crawling strategy, we sourced pairs of images with corresponding alt-texts.
An issue for anti-AI people, as seen on Bluesky, is that they're often "insisting you write alt text for all images" people as well. But this is probably the main use for alt text at this point, so they're essentially doing annotation work for free.
It's fine if you want to, but I think they should consider that basically nobody is reading it. If it was important for society, photo apps would prompt you to embed it in the image like EXIF.
Computer vision is getting good enough to generate it; it has to be, because real-world objects don't have alt text.
I actually use Claude to generate the first draft of most of my alt text, but I still do a manual review of it because LLMs usually don't have enough contents to fully understand the message I'm trying to convey with an image: https://simonwillison.net/2025/Mar/2/accessibility-and-gen-a...
Why would photo apps do what's "important for society"?
Annotating photos takes time/effort, and I could totally imagine photo apps being resistant to prompting their users for that, some of which would undoubtedly find it annoying, and many more confusing.
Yet I don't think that one can conclude from that that annotations aren't helpful/important to vision impaired users (at least until very recently, i.e. before the widespread availability of high quality automatic image annotations).
In other words, the primary user base of photo editors isn't the set of people that would most benefit from it, which is probably why we started seeing "alt text nudging" first appear on social media, which has both producer and consumer in mind (at least more than photo editors).
> Why would photo apps do what's "important for society"?
One would hope they're responsive to user demands. I should say Lightroom does have an alt text field, but like phone camera apps don't.
Apple is genuinely obsessed with accessibility (but bad at social media) and I think has never once advocated for people to describe their photos to each other.
> An issue for anti-AI people, as seen on Bluesky, is that they're often "insisting you write alt text for all images" people as well. But this is probably the main use for alt text at this point, so they're essentially doing annotation work for free.
How did you come to the conclusion that those two groups overlap so significantly?
This is a well known fact. A bunch of AI researchers tried to migrate to the platform from Twitter but got a ton of hate and death threats from other users so they went back. Bluesky has a pretty strong anti-AI bias and the community of folks talking about it despite that is very small.
So you found a couple people expressing this conflicting view and assumed it applies to a larger group? Doesn’t sound very reliable to me but I see this all the time and it makes sense if you look at it as a mechanism to explain the world .
Gotta polish that fig-leaf to hide Apple's real stance towards user privacy: arstechnica.com/tech-policy/2023/12/apple-admits-to-secretly-giving-governments-push-notification-data/
> Apple has since confirmed in a statement provided to Ars that the US federal government "prohibited" the company "from sharing any information,"
> We do not use our users’ private personal data or user interactions when training our foundation models. Additionally, we take steps to apply filters to remove certain categories of personally identifiable information and to exclude profanity and unsafe material.
> Further, we continue to follow best practices for ethical web crawling, including following widely-adopted robots.txt protocols to allow web publishers to opt out of their content being used to train Apple’s generative foundation models. Web publishers have fine-grained controls over which pages Applebot can see and how they are used while still appearing in search results within Siri and Spotlight.
Respect.