Corellium Launching New Initiative to Hold Apple Accountable Over CSAM Detection Security and Privacy Claims

[Playfoot]

Interesting, Apple announced today they appealed a copyright case they lost against Corellium....Lost last year. So timing is a little odd.

Much could be read into this, and of course we mere mortals do not have the information. However, it could be viewed as an effort by Apple to quash any true third party research/investigation into the CSAM Hashgate. Is this an example of Apple publicly stating desire for third party contribution, but in reality doesn't....

 

[usagora]

At this point, I can't remember. However, I must assume there was something you wrote that made me think there was a legal question involved...Or I mistakenly clicked on your message.

No, not nefarious, just questioning why. In my opinion Apple has needlessly burned privacy capital, Apple has "needed hard" to create a system that could be used for State security, Apple has fumbled the launch and the PR. Why all of this when it was NOT required by law? Some have pointed out a greater and better present such as expanded E2E. However, no one is in possession of a crystal ball.

Well if it's not nefarious, then who cares? And there you go again with implying something shouldn't be done simply because it's not required by law. If you saw an elderly lady fall on the ground, are you required by law to help her? No, but you (hopefully) still do so because you're a decent human being. And that's just an analogy, so please don't reply back that I'm comparing CSAM detection to an elderly lady falling down - I'm simply making the point that both individuals and companies often do things that aren't required by law because it's simply the decent and right thing to do.

The "why" in terms of the change in the scanning process has already been answered by Apple, and ironically it's the very thing that people are acting like this violates: greater privacy. The "why" in terms of why they scan at all (whether on the cloud or on the device) is because decent companies are proactive about preventing the spread of CSAM, whether the law requires them to or not.

I don't think Apple fumbled anything. Yes, I know Craig F. said some self-deprecating things about their rollout, but that's just him being a diplomatic PR person in response to all the paranoia and confusion caused either by people not reading carefully or purposely reading between the lines to create drama where there is none. In his mind he's probably rolling his eyes at all the ridiculous misinterpretations and conspiracy theories surrounding a pretty simple concept as they explained it.

 

[Playfoot]

Interesting . . .

Apple’s child-abuse scanner has serious flaw, researchers say

Researchers have produced a collision in Apple’s NeuralHash system

www.msn.com

 

[Playfoot]

Well if it's not nefarious, then who cares? And there you go again with implying something shouldn't be done simply because it's not required by law. If you saw an elderly lady fall on the ground, are you required by law to help her? No, but you (hopefully) still do so because you're a decent human being. And that's just an analogy, so please don't reply back that I'm comparing CSAM detection to an elderly lady falling down - I'm simply making the point that both individuals and companies often do things that aren't required by law because it's simply the decent and right thing to do.

The "why" in terms of the change in the scanning process has already been answered by Apple, and ironically it's the very thing that people are acting like this violates: greater privacy. The "why" in terms of why they scan at all (whether on the cloud or on the device) is because decent companies are proactive about preventing the spread of CSAM, whether the law requires them to or not.

I don't think Apple fumbled anything. Yes, I know Craig F. said some self-deprecating things about their rollout, but that's just him being a diplomatic PR person in response to all the paranoia and confusion caused either by people not reading carefully or purposely reading between the lines to create drama where there is none. In his mind he's probably rolling his eyes at all the ridiculous misinterpretations and conspiracy theories surrounding a pretty simple concept as they explained it.

Well, we agree to disagree.

 

[JMacHack]

Because it is sufficiently repressive that we can't meaningfully vote against said repression, and can't emigrate to anywhere with a meaningful democracy.

1. Germany, apparently. They asked Apple to remove this “feature”.

2. Frederick Douglass said “Democracy works on three boxes, the soapbox, the ballot box, the jury box, and the cartridge box.”

At the risk of being called an extremist, or in fact convicted on “wrongthink”, I believe we’re at jury box, and may have to soon resort to cartridge box.

 

[MacBH928]

These are all excluded middle fallacies to support an argument from adverse consequences positions.

Heck, two of them are total non sequiturs.

-If the reason that there are lower interest rates is if 1/5 are out of work how does it follow more people will take loans? Sure as bloody well didn't work that way during the Long Depression or Great Depression.

-If a discount is made and nobody wants or knows about the product how does it follow sales will rise?

I am sorry, your wording is confusing me but I have already stated we do not know for sure what is going to happen but we can come close. There are a lot of other factors, and we can't tell the future.

-Apple could go crazy and be worse than facebook

-Tim Cook might feel bad about CSAM reaction and decide to turn all Apple software open source.

who knows?! We can only make a guess-estimate from past historic instances.

 

[JMacHack]

EX.
-If we let all criminals in jail out -> crime incidents will increase
-If we lower interest rate -> more people will take loans
-If we make a discount -> our sales will rise

To be pedantic, I think you’re casting too wide a net with these examples.

1. Criminal recidivism depends on many circumstances beyond just being convinced of a crime. In the case of the war on drugs we’ve seen that harsh punishments do not deter drug use for example.

2. While broadly lowering interest rates would encourage people to take loans for larger amounts of money. It only works if there’s a baseline amount of people earning just enough to be denied a loan based on interest rates alone. For example in a highly stratified income society, many could simply not afford loans at all because they live hand to mouth.

3. This implies the only factor in purchasing a product is price. Many factors go in to individual purchase choices other than price. For example look at the focus group reactions to the Pontiac Aztek. They said “I would not want one if given to me.”

 

[WaffleBoi_15]

It's too late for me, I had the AppleOne Family Premium subscription and used over 1TB of my storage space with photos & documents. I've removed everything from iCloud and disabled iCloud Drive and backups on all my devices, and downgraded my plan to the lowest.
Such a bummer, I've heavily invested in the Apple ecosystem. All of my family's devices are Apple products (there are 7 of us) but iCloud will be restricted to Calander & contacts for us going forward.

I've lost my faith in Apple, after 30 years for me....

I hope you eventually learn there was no point to this.

 

[Maximara]

I think that's his point: they don't really care. If they did, there would at the very least be effective enforcement against the catholic church, the boy scouts, and so on, plus legislation to punch though the many corporate veils the catholic church hides behind to ensure that no organisation is ever liable for anything if it has any assets.

The problem is that people weren't reporting these abuses (or if they did they were ignored by the local law enforcers). With regards to the Catholic Church, it is already getting hit in the pocket book: The Clergy Abuse Crisis Has Cost The Catholic Church $3 Billion and has gotten to the point Pope outlaws church-related sex abuse

If they were a bit more enthusiastic, they'd push for a WTO rule allowing (or even requiring) countries to require importers to prove that they don't use children, slaves, etc..

The WTO is a joke. If they were serious they would crack down on the slave-like labor that goes on in places like Africa, China, or even here in the US (estimates are around 7,500 foreign nationals and 400,000 Americans). Welcome to the real world.

 

[Maximara]

Your logical fallacy is the fallacy fallacy

You presumed that because a claim has been poorly argued, or a fallacy has been made, that the claim itself must be wrong.

yourlogicalfallacyis.com

@MacBH928 rattled off a couple of quick examples with fairly loose wording that superficially invoked "excluded middle" - but, heck, this is an internet forum not a prestigious academic journal where authors might be expected to provide a reference for every last assertion made or implied.

In your "debunking" you have:

(a) Taken "->" to mean "it is inevitable that" rather than "it is likely that". Almost any discussion about social/economic issues involves likelihood rather than deterministic cause and effect.

Actually I was reading "->" in the manner it was (and still is) used in the college level math textbooks in the late 80's and early 90's (I have a minor in mathematics). The formal meaning of "->" was (and still is) "then", "hence", "therefore", or any other words in the same spirit.

What you are describing is the "fat arrow" (=>) which does indeed mean "implies" which was used in none of the statements. MacBH928 used the wrong arrow (->) in the statements if that is what they ment.

(b) Assumed that the "missing link" is unlikely or contentious, and needs spelling out - it really isn't an extraordinary claim that lower interest rates might encourage more people to take out loans, or that released criminals are more likely to re-offend than the average person.

The formal meaning of "->" was (and still is) "then", "hence", "therefore", or any other words in the same spirit. It does not mean "might". That again is the "fat arrow" (=>)

"-If we lower interest rate -> more people will take loans" means
-If we lower interest rate then more people will take loans
-If we lower interest rate therefore more people will take loans
Hence doesn't really make sense.

It's highly unlikely that a seller would offer a discount without advertising it.

You would be surprised. Everybody knew about the blue light specials at K-mart but no one knew what they were on. Never mind that many such advertisements tend to be what are known as loss leaders. Never mind that some consumers caught on to the 'discount an artificially marked up item' ploy.

 

[Maximara]

Interesting . . .

Apple’s child-abuse scanner has serious flaw, researchers say

Researchers have produced a collision in Apple’s NeuralHash system

www.msn.com

"Actually generating that alert would require access to the NCMEC hash database, generating more than 30 colliding images, and then smuggling all of them onto the target’s phone. Even then, it would only generate an alert to Apple and NCMEC, which would easily identify the images as false positives"

 

[theluggage]

Actually I was reading "->" in the manner it was (and still is) used in the college level math textbooks in the late 80's and early 90's

...and that's the problem: you assume that someone taking part in an informal discussion is using a particular form of formal notation with which you are familiar (other meanings of "->" are available - to most people it's just a way of typing an arrow). Then you're using that to dismiss the entire argument without even explaining your assumptions about the notation.

The formal meaning of "->" was (and still is) "then", "hence", "therefore", or any other words in the same spirit. It does not mean "might". That again is the "fat arrow" (=>)

Straw man/Appeal to the dictionary: nobody is questioning the correct definition and usage of the formal logic notation you are describing. The problem was your unfounded and unstated presumption that the poster was trying to use that particular notation, and that no other interpretation is possible.

Actually "=>" is a function definition in Javascript and and "->" is pointer dereferencing in C... except that if I actually look and think about what you have written and consider the context, I see that you are evidently not trying to write a Javascript or C program, so those definition wouldn't apply. You're assuming that the poster was trying to write a formal proof using a particular notation. Even so, you could have just called them out on their misuse of "->".

MacBH928 used the wrong arrow (->) in the statements if that is what they ment.

...perfectly correct, when stated with that added condition.

Everybody knew about the blue light specials at K-mart but no one knew what they were on.

So they were advertising the discounts. People would visit K-Mart to see what was on offer. People would likely buy more of the discounted items. The fact that K-Mart's objective might have been to attract more people to the shop and increase sales of other items doesn't change that.

Even then, it would only generate an alert to Apple and NCMEC, which would easily identify the images as false positives"

Do you mean
"-> NCMEC would identify the images as false positives"
or
"=> NCMEC would identify the images as false positives"
?

 

[Maximara]

Straw man/Appeal to the dictionary: nobody is questioning the correct definition and usage of the formal logic notation you are describing. The problem was your unfounded and unstated presumption that the poster was trying to use that particular notation, and that no other interpretation is possible.

I read "->" as it generally is used in the material I am familiar with. Instead of using actual words they used a symbol which has a particular meaning.

Actually "=>" is a function definition in Javascript and and "->" is pointer dereferencing in C... except that if I actually look and think about what you have written and consider the context, I see that you are evidently not trying to write a Javascript or C program, so those definition wouldn't apply. You're assuming that the poster was trying to write a formal proof using a particular notation. Even so, you could have just called them out on their misuse of "->".

They stated "This is no logical fallacy. This is truth to logic"

That sets the context in which "->" is read (from Contrary Logic):
"p⇒q means 'IF p is true THEN q is true."
From a 1970s book on logic:
Implication p -> q “if p then q”

Do you mean
"-> NCMEC would identify the images as false positives"
or
"=> NCMEC would identify the images as false positives"
?

You are being silly and dodging the issue. Instead of using actual words they went lazy and tried to use a system that in the context of both logic and math means "then". Heck even your counter example means "then" not "might".

None of this counters by lede in: "These are all excluded middle fallacies to support an argument from adverse consequences positions."

 

[Maximara]

To be pedantic, I think you’re casting too wide a net with these examples.'=

1. Criminal recidivism depends on many circumstances beyond just being convinced of a crime. In the case of the war on drugs we’ve seen that harsh punishments do not deter drug use for example.

2. While broadly lowering interest rates would encourage people to take loans for larger amounts of money. It only works if there’s a baseline amount of people earning just enough to be denied a loan based on interest rates alone. For example in a highly stratified income society, many could simply not afford loans at all because they live hand to mouth.

3. This implies the only factor in purchasing a product is price. Many factors go in to individual purchase choices other than price. For example look at the focus group reactions to the Pontiac Aztek. They said “I would not want one if given to me.”

Right, that is why I called the examples effectively excluded middle fallacies with some adverse consequences, and non sequtor thrown in for good measure.

Heck, looking back at them I also see post hoc, ergo propter hoc, observational selection, confusion of correlation and causation, and half-truth fallacies as well.

That is a lot of oopies to hang the hat on fallacy of fallacy on especially when we are dealing with something so complex.

 

[theluggage]

I read "->" as it generally is used in the material I am familiar with. Instead of using actual words they used a symbol which has a particular meaning.

To repeat: I'm not challenging what the notation means. I'm challenging your assumption that the original poster intended to use that specific meaning of '->'.

They stated "This is no logical fallacy. This is truth to logic"

OK, now you are starting to improve your argument by justifying your decision to hold the post to a higher level of logical precision.

"p⇒q means 'IF p is true THEN q is true."

...then we're back to you irrelevantly re-stating facts that aren't in contention.

You're also contradicting yourself BTW:

What you are describing is the "fat arrow" (=>) which does indeed mean "implies" which was used in none of the statements. MacBH928 used the wrong arrow (->) in the statements if that is what they ment.

According to this - https://en.wikipedia.org/wiki/List_of_logic_symbols - '=>' and '->' are synonymous. Neither of then mean "if then maybe b". You're really talking about propositional calculus which deals with propositions that are, by definition, either true or false.

None of this counters by lede in: "These are all excluded middle fallacies to support an argument from adverse consequences positions."

Because you didn't go on to support that assertion and instead focussed on the alleged misuse of "->" which - right or wrong - has little or nothing to do with "excluded middle fallacies".

In fact, the "excluded middle fallacy" here is that you're misapplying the "excluded middle" rule for logical propositions (if A is not true than A must be false) to discussions about risks and likelihoods which are outside the domain of propositional logic (but not logic in general).

The "fallacy fallacy" applied because you are attacking the language and notation of a statement (which, it is quite true, was badly argued) and ignoring the claim that "slippery slope" is not a fallacy if there's a plausible justification for the risk that doing 'A' may lead to "B' and 'C'. Go read the "fallacy fallacy" explanation again - it doesn't mean that no fallacy was committed.

 

[Maximara]

According to this - https://en.wikipedia.org/wiki/List_of_logic_symbols - '=>' and '->' are synonymous. Neither of then mean "if then maybe b". You're really talking about propositional calculus which deals with propositions that are, by definition, either true or false.

Stop right there. As I pointed out before in the "ssd swap - high usage of Terabytes Written" thread:

For those who, for some silly reason, want to point to wikipedia read Wikipedia is not a reliable source: "Because it can be edited by anyone at any time, any information it contains at a particular time could be vandalism, a work in progress, or just plain wrong. (...) Wikipedia is a volunteer-run project, it cannot constantly monitor every contribution. There are many errors that remain unnoticed for hours, days, weeks, months, or even years. Therefore, Wikipedia should not be considered a definitive source in and of itself"

Even worse the only reference in the relevant section on the wikipedia page (at the time of writing this response) has the circular tag ie the reference to back up the statement is effectively taken from wikipedia or a similar source.

I should point out that even if there was a non wiki reference it may not be valid. See the "Vampire connection, phony reference" on the Red Riding Hood talk page for an example of someone catching a made up reference.

So your only counter argument is a source that itself says its articles are not reliable enough to be used in its own articles! More over relevant section of the page uses a source of similar quality as a reference!

Wikipedia is good for a quick look up but I would never use it directly and I would always cross check any reference to make sure it actually said what the contributor claims it says.

By contrast the Contrary Logic I used is produced by the University of Cambridge which even wikipedia would consider a reliable source.

"Implies" is the connective in propositional calculus which has the meaning "if

is true, then

is also true." - Eric Weisstein at Wolfram Research This is different from its layman use of "strongly suggest the truth or existence of" (Oxford)

This reminds of a comment from Doctor Who of all places:
DOCTOR: All elephants are pink. Nellie is an elephant, therefore Nellie is pink. Logical?
DAVROS: Perfectly.
DOCTOR: You know what a human would say to that?
DAVROS: What?
TYSSAN: Elephants aren't pink.
DAVROS: Bah. Humans do not understand logic.
ROMANA: They're not slaves to it like the Daleks or the Movellans.

In short, Elephants are pink ∧ Nellie is an elephant => Nellie is pink may be perfectly logical...but it is also totally nonsensical.

 

[theluggage]

Even worse the only reference in the relevant section on the wikipedia page (at the time of writing this responce) has the circular tag ie the reference to back up the statement is effectively taken from wikipedia or a similar source.

https://stanford.edu/~shervine/teaching/cs-221/cheatsheet-logic-models (-> means implication/if-then)

https://www.maths.usyd.edu.au/MATH1901/r/PDF/cheat-sheet.pdf ( => means implication... -> used in the context of functions)

...that's two major universities using "->" and "=>" interchangeably to mean "implies" not to mention:

That sets the context in which "->" is read (from Contrary Logic):
"p⇒q means 'IF p is true THEN q is true."
From a 1970s book on logic:
Implication p -> q “if p then q”

...which is entirely consistent with what Wikipedia and the other references say. I was actually pointing out that you were contradicting yourself, since you also said:

The formal meaning of "->" was (and still is) "then", "hence", "therefore", or any other words in the same spirit.

What you are describing is the "fat arrow" (=>) which does indeed mean "implies" which was used in none of the statements. MacBH928 used the wrong arrow (->) in the statements if that is what they ment.

As is pretty clear from all the references above, both '->' and '=>' mean "imp-lies" and, in propositional logic "implies" and "then/therefore/hence" are synonymous because in that context, by definition all propositions are either true or false and there's no concept of 'maybe'.

 

[Maximara]

As is pretty clear from all the references above, both '->' and '=>' mean "imp-lies" and, in propositional logic "implies" and "then/therefore/hence" are synonymous because in that context, by definition all propositions are either true or false and there's no concept of 'maybe'.

Again "Implies" is the connective in propositional calculus which has the meaning "if

is true, then

is also true." - Eric Weisstein at Wolfram Research

This is different from its layman use of "strongly suggest the truth or existence of" (Oxford). You are using the later definition while I am using the former (which is how it is used in logic).

"The statement “p implies q” means that if p is true, then q must also be true. The statement “p implies q” is also written “if p then q” or sometimes “q if p.” Statement p is called the premise of the implication and q is called the conclusion" - Northern Illinois University Department of Mathematical Sciences, MATH 101 Spring 2007, Implications handout.

 

[theluggage]

Again "Implies" is the connective in propositional calculus which has the meaning "if

is true, then

is also true." - Eric Weisstein at Wolfram Research

Yes, that's effectively what I just said, it's what wikipedia said, and it is consistent with the two other references I linked to (which use either '=>' or '->' to mean "Implies") when you rejected the wikipedia link - the only thing it is inconsistent with is your earlier statement:

The formal meaning of "->" was (and still is) "then", "hence", "therefore", or any other words in the same spirit.

What you are describing is the "fat arrow" (=>) which does indeed mean "implies" which was used in none of the statements. MacBH928 used the wrong arrow (->) in the statements if that is what they ment.

So you were saying that '->' means then/hence/therefore and '=>' means "implies" in the colloquial 'strongly suggests' sense (MacBH928 using 'implies' in the prop logic sense would actually have made things worse).

This is different from its layman use of "strongly suggest the truth or existence of" (Oxford).

Yes.

You are using the later definition while I am using the former (which is how it is used in logic).

Where? Pretty sure I've only used "implies" in the previous post where I was saying it was equivalent to if/then (unless you count using it in a different context).

...but your whole problem remains that you're trying to apply formal logic standards to an informal discussion, just because another poster used the word "logic" and a '->' symbol sloppily in an argument. We're not writing math proofs here. Propositional logic dealing with true/false statements isn't even useful in discussing complex real-world situations where everything is a risk assessment - unless you want to trace every bit of math back to its fundamental particles.

 

[Maximara]

Where? Pretty sure I've only used "implies" in the previous post where I was saying it was equivalent to if/then (unless you count using it in a different context).

Here:
"(a) Taken "->" to mean "it is inevitable that" rather than "it is likely that". Almost any discussion about social/economic issues involves likelihood rather than deterministic cause and effect."

"Implies" is the connective in propositional calculus which has the meaning "if

is true, then

is also true." - Eric Weisstein at Wolfram Research. This is different from its layman use of "strongly suggest the truth or existence of" (Oxford)

There is the proof that as I said "You are using the later definition while I am using the former (which is how it is used in logic)."

Using the terminology in the Informal Fallacies part 2 handout for Math 101, Spring 2007 handout, Department of Mathematical Sciences, Northern Illinois University this would fall under Equivocation Fallacy (Use of expressions of double meaning in order to mislead) is it was intentional which it likely wasn't. I would like to point out that JMacHack in post #157 also pointed out the excluded middle aspect of the examples.

Regarding Equivocation Fallacy one sees the same nonsense with the common misuse of "myth" to mean 'totally untrue story" effectively put the stories of the Greek gods or the claim sound films didn't exist before 1927 on par with the story Columbus sailing West to prove the Earth was round.

"Conspiracy Theory" has the same problem effectively equating the idea the American mafia made Hoffa "disappear" on par with Area 51 testing captured alien spacecraft. It has gotten so bad that Parish, Jane (Editor), Martin Parker (Editor) (2001) The Age of Anxiety: Conspiracy Theory and the Human Sciences, Wiley-Blackwell ISBN: 978-0-631-23168-4 tried to coin the term "Theory of Conspiracy" to separate the two.

Basically the "◊" was left out of the statements resulting in over generalizations.

To get this back on track (we have gone down this logical rabbit hole way past the point of relevance) at least Apple told people what it was going to do. That is a lot more than the Lower Merion School District
or Aaron's Inc did - they did it in secret and if they hadn't gotten caught they likely would still be doing their invasion of privacy nonsense.

One possibility is this thing is based on complying with laws on the books (18 U.S. Code § 2258 and its four sub laws here in the US) allowing Apple to take a high moral ground on government efforts to open up their iPhones.

"But if we allow sideloading they will be able to bypass this measure. Think of the children that will be harmed!" and to their customers they will point to those very same laws and likely say "But the law effectively requires us to do this - to protect the children."

Then they break out the popcorn and watch the political Catch-22 free-for-all.

 

[theluggage]

Here:
"(a) Taken "->" to mean "it is inevitable that" rather than "it is likely that". Almost any discussion about social/economic issues involves likelihood rather than deterministic cause and effect."

Yes, because your initial argument relied on assuming one very specific meaning of '->' in formal math/logic while ignoring other possible meanings that were actually more plausible in context (and the original poster has long since confirmed that they were talking about likelihood, not absolutes). You hadn't even stated your assumption of formal logic notation at that point.

...and you're still compounding your appeal-to-the-dictionary fallacy - I referred to the OP's use of '->' - you choose to interpret that as "implies" and then look up "implies" in the dictionary... The whole point of my argument is that you weren't justified in interpreting the original post as formal propositional logic.

...and you still haven't explained what you meant by:

The formal meaning of "->" was (and still is) "then", "hence", "therefore", or any other words in the same spirit.

What you are describing is the "fat arrow" (=>) which does indeed mean "implies" which was used in none of the statements. MacBH928 used the wrong arrow (->) in the statements if that is what they ment.

...which appears nonsensical since we've established that 'A => B', 'A -> B', 'A implies B', 'if A then B" are - in the context of formal logic notation - synonymous. MacBH928's post was sloppily argued - that is not in dispute - but it certainly wouldn't have been 'fixed' by changing '->' to '=>'.

Mathematical Sciences, Northern Illinois University this would fall under Equivocation Fallacy (Use of expressions of double meaning in order to mislead) is it was intentional which it likely wasn't.

...but if it wasn't intentional, and you base your rebuttal on one specific meaning without permitting the alternatives, you're committing an Appeal to the Dictionary fallacy, which is what I'm claiming here.

I would like to point out that JMacHack in post #157 also pointed out the excluded middle aspect of the examples.

...but that post still hinges on the presumption that the OP intended a specific meaning of '->' and disregarding that sometimes an arrow is just an arrow (Pro tip: never go to a seminar on social sciences if you are inclined to think of boxes connected by arrows as representing nodes and edges on a graph with some formal significance - you'll go mad )

Also - careful with that "excluded middle" stuff - "excluded middle fallacy" is misuse of the "law of excluded middle" which is about true/false propositions. If an argument is not a true/false proposition then the fallacy is attacking it "because excluded middle". Again, it all hinges on your assumption that the OP intended '->' in the propositional logic sense, and they have already clarified that they didn't - the elephant is not pink.

at least Apple told people what it was going to do

Absolutely - and the positive side of that is that a lot of independent experts are now scrutinising what they've claimed... Unlike a random person's posts on web forums, Apple's releases should be assumed to have been carefully written and thoroughly reviewed before released, and should be held to a very high standard of technical accuracy and

The most detailed papers are too technical for anybody but crypto/image recognition experts, so I'm not even going to try, but you might want to attack this one with a dictionary: https://www.apple.com/child-safety/pdf/CSAM_Detection_Technical_Summary.pdf - e.g. the way they start with "unique number specific to that image" on page 4, and work through "nearly identical" (a bit like "only slightly pregnant") until by page 5 it's "identical and visually similar images result in the same hash". ("similar" has a strict true/false geometrical meaning - e.g. 'similar triangles' - but 'visually similar' introduces an unknown degree of uncertainty to that).

Certainly if you look at these threads you'll see lots of people who have read Apple's press releases have come away with the message "it's not scanning my photos because it just compares hashes" which is a great example of an equivocation fallacy (conflating "NeuralHash perceptual hashes" - an image recognition technique - with cryptographic hashes, a security technique - both "hashes" in the broadest sense but hugely different in the details and implications{colloquial meaning} ).

 

[Maximara]

Absolutely - and the positive side of that is that a lot of independent experts are now scrutinising what they've claimed... Unlike a random person's posts on web forums, Apple's releases should be assumed to have been carefully written and thoroughly reviewed before released, and should be held to a very high standard of technical accuracy and

The most detailed papers are too technical for anybody but crypto/image recognition experts, so I'm not even going to try, but you might want to attack this one with a dictionary: https://www.apple.com/child-safety/pdf/CSAM_Detection_Technical_Summary.pdf - e.g. the way they start with "unique number specific to that image" on page 4, and work through "nearly identical" (a bit like "only slightly pregnant") until by page 5 it's "identical and visually similar images result in the same hash". ("similar" has a strict true/false geometrical meaning - e.g. 'similar triangles' - but 'visually similar' introduces an unknown degree of uncertainty to that).

Reminds me of Burke in Connections who talked about how technical terminology had gotten to the point that only other experts can really understand what is going on. My later mother, an editor for Ohio State University, had this problem when what was supposedly written for public consumption...but was written as if it was for other mathmations.

My own field, anthropology, has a similar problems with the term "culture" - without context there is no way to tell what it means. Your example, "nearly identical" actually has this issue: "Only another image that appears nearly identical can produce the same number; for example, images that differ in size or transcoded quality will still have the same NeuralHash value."

How Machines Learn is a quick explanation of the process of how these bots work. It also indicates why some videos do things like speeding up the audio or flipping the image of movies - its an attempt to avoid the copyright bots.

As for Apple explaining things to the public...you have a public that doesn't understand the difference between a code and cypher.