Live data from Hacker News

Google Drive flags file only containing “1” for copyright infringement

twitter.com

161–170 of 464 posts

Re: Google Drive flags file only containing “1” for copyright infringement

#161

"A review cannot be requested for this restriction" ML enforcing rules is bad enough, but not allowing false positives to be corrected is ridiculous. This is why I would never consider g-suite for any business application. Otoh, I think there is a legitimate business to be made helping small businesses and individuals secure themselves against arbitrary behavior from big tech. This kind of thing can have serious cons…

Given that going viral on Twitter is the only functional help desk for such situations, “Tweet storm as a service” would be a promising proposition.

Companies can say that they won't have an actual human look into your issues, but there's a person that handles letters that might turn into lawsuits and other legal issues, and you can always ensure that they get contacted instead. Certified mail is pretty quick to write and costs like $3 to send, and you can look up the company's agent for service of process.

It might not do anything - writing a letter that sufficiently implies that you are actually collecting documentation and preparing for a lawsuit is an actual skill, and your demands may be unreasonable for them to handle - but you will get an actual factual human being to at least start reading what you've written. And as a bonus, their KPI is in terms of "number of incidents per year" rather than "number of resolved tickets per day".

Re: Google Drive flags file only containing “1” for copyright infringement

#162

Earlier quoted context omitted.

That's ridiculous. Storing social security numbers is necessary for lots of businesses. Imagine your filing cabinet not letting you file employment forms with a SS# on them. It's the sticky note password problem all over again...

> That's ridiculous. Storing social security numbers is necessary for lots of businesses. They should absolutely not be stored in a GSuite document. SSNs should be treated more securely than credit card numbers.

Which is a problem of its own, since they're effectively usernames.

Secure usernames that have no corresponding password is already an oxymoron; that's what credit card numbers used to be, hence the introduction of the CVV, 3DSecure and so on; but at least a credit card can be blocked with relative ease. But SSNs are secure unchangeable usernames, which makes even less sense.

Do any countries other than the US have such an abomination, where you can figure out the SSN of someone and ruin their life?

Re: Google Drive flags file only containing “1” for copyright infringement

#163

Earlier quoted context omitted.

Since they went wide with "intellectual property rights" there, the references to health and safety are probably more in the realm of trademark and maybe patent... think counterfeit drugs. You can probably gin up a copyright example from, I dunno, the DRM system on some medical device or something, though that's obviously not the real focus of their copyright enforcement work.

But drug safety is not an issue of copyright, but of physical control of medications. You don't need to break the copyright of a drug to create and distribute fake medication.

But you might be able to trick more people to buy your fake medication if you put another company's logo on it. I'd hate to think that a country's procedures for stopping copyright infringement are so efficient that they are the optimal route for preventing the distribution of fake medication (even making trademark law pointless), but I must admit that comparing two images is simpler than comparing the composition of two medications. Corporate needs you to find the difference between these two approaches.

Re: Google Drive flags file only containing “1” for copyright infringement

#164
post #58

I have a pet theory that all of these recent Google bloopers could be explained easily if you start from the assumption that Google internal incentives promote efforts to cut costs such as storage. "Garbage" docs, inactive email accounts, less search results etc can all be reasonably explained by a desire to not spend money on storage for "low value" data (i.e. data that is unlikely to be accessed in a way that trans…

> explained by a desire to not spend money on storage for "low value" data

I think this "just happens" when you reach a certain scale. For example, I was looking at our reporting at my job the other day and realized for 10k people a day, something wasn't working with our emails. We decided it was a "low priority" because it's "only" 10k people when we send over 30 million emails a day. I'm sure those 10,000 people (a small city's worth!) don't feel that way.

Re: Google Drive flags file only containing “1” for copyright infringement

#165
post #77

Earlier quoted context omitted.

> "tiny" scenarios which finally does Google in Danger of tiny scenarios - I expect that Google like any BigCo will try files containing only 2, 3, 7 - no bug, ok, and then push the fix like this if (l = read(file) ; l == "1") ... else

Given it's almost certainly a ruleset generated by an ml agent, it's more likely to be a change in the training data.

Or just don't flag any file below 100 bytes.

Re: Google Drive flags file only containing “1” for copyright infringement

#166

Earlier quoted context omitted.

That's ridiculous. Storing social security numbers is necessary for lots of businesses. Imagine your filing cabinet not letting you file employment forms with a SS# on them. It's the sticky note password problem all over again...

> That's ridiculous. Storing social security numbers is necessary for lots of businesses. They should absolutely not be stored in a GSuite document. SSNs should be treated more securely than credit card numbers.

Should, yes, but in practice not really. I'm talking more about employee information.

You need it for tax forms, background checks, citizenship queries, sometimes bank information, etc.

So your options are:

1. Store them locally on a computer. Typically on some old windows 7 machine in the corner that hasn't been updated in some time.

2. Store documents physically. Which will either be scanned onto random computers belonging to whoever needs them to be sent through probably insecure mail servers.

Or worse, your boss taking a picture of your form and sending it to people that way, leaving the form on their phone.

3. Some other online storage like whatever M$ is offering

4. Use google and somehow store SS#'s somewhere less secure, or obfuscate them in a way no one but a few people will understand and hope they don't block any other files you upload.

Businesses have been deciding how to manage these things since the start that work best for them. Having google force you into procedures that might not work for your use case is annoying at best. And they obviously don't know best if they have issues like in OP's post.

It's like they take away your gun so you can't shoot yourself in the foot, then fires it at things it thinks are problems hoping not to hit your feet.

And what's a few toes to a company the size of google?

Re: Google Drive flags file only containing “1” for copyright infringement

#167

Earlier quoted context omitted.

> That's ridiculous. Storing social security numbers is necessary for lots of businesses. They should absolutely not be stored in a GSuite document. SSNs should be treated more securely than credit card numbers.

Which is a problem of its own, since they're effectively usernames. Secure usernames that have no corresponding password is already an oxymoron; that's what credit card numbers used to be, hence the introduction of the CVV, 3DSecure and so on; but at least a credit card can be blocked with relative ease. But SSNs are secure unchangeable usernames, which makes even less sense. Do any countries other than the US have s…

Not that I know of. It really is a horrible system. But it's not one I want Google trying to solve.

Re: Google Drive flags file only containing “1” for copyright infringement

#168
post #64
post #22

Earlier quoted context omitted.

Reminds me of that 70s IBM presentation quote that surfaced recently, "A computer can never be held accountable, therefore a computer must never make a management decision". It's one thing to have a computer flag issues, another to make it responsible for taking action and in this case, making a decision final. Google continues to set poor examples with irresponsible implementations of machine learning. With no accou…

I think that quote should be flipped. "A computer can never be held accountable, therefore a computer must always make management decisions."

That misunderstanding could explain the last several decades of corporate management.

Re: Google Drive flags file only containing “1” for copyright infringement

#169
post #7

"A review cannot be requested for this restriction" ML enforcing rules is bad enough, but not allowing false positives to be corrected is ridiculous. This is why I would never consider g-suite for any business application. Otoh, I think there is a legitimate business to be made helping small businesses and individuals secure themselves against arbitrary behavior from big tech. This kind of thing can have serious cons…

How long until cloud providers are forced to scan block devices, so you can't even self-host a file erroneously flagged by the ML? (Actually, how long until this is integrated with the "security engine" in CPUs, so you can't even hide with encryption? Everyone is talking about on-device voice-recognition and CSAM recognition, but no doubt the copyright lobby is going to push hard for this as well.)

If Amazon didn't offer EBS block device encryption (with a key that's ostensibly only accessible to the customer), then customers would just use full disk encryption instead.

Re: Google Drive flags file only containing “1” for copyright infringement

#170
Google Drive also offers client side encryption, which would make this scanning ineffectual: https://flowcrypt.com/blog/article/2021-06-14-google-workspa...

So as long as you have a ton of money and are a corporation your privacy should be just fine

Post reply on HN