Response to EFF complaint about student data privacy
googleforeducation.blogspot.com
googleforeducation.blogspot.com
> There are no ads in these Core Services, and student data in these services is not used for advertising purposes.
This isn't in the original EFF complaint, EFF specifically says (at least in the press release [1]) that Google doesn't advertise:
EFF:
>This allows Google to track, store on its servers, and data mine for non-advertising purposes
Google then says:
>In addition, our systems compile data aggregated from millions of users of Chrome Sync and, after completely removing information about individual users, we use this data to holistically improve the services we provide. For example if data shows that millions of people are visiting a webpage that is broken, that site would be moved lower in the search results. This is not connected to any specific person nor is it used to analyze student behaviors.
Which is exactly what the EFF's complaint is about!
EFF:
>Google’s practices fly in the face of commitments made when it signed the Student Privacy Pledge, a legally enforceable document whereby companies promise to refrain from collecting, using, or sharing students’ personal information except when needed for legitimate educational purposes or if parents provide permission.
so imho Google just confirms what the EFF said
[1] https://www.eff.org/press/releases/google-deceptively-tracks...
LIZ FIGUEROA, (D) State Senator, CA, 1998-06: We walk into this room, and it’s myself
and two of my staff— my chief of staff and one of my attorneys. And across from us
was Larry, Sergey, and their attorney.
All of a sudden, Sergey started talking to me. He said, “Senator, how would you feel
if a robot went into your home and read your diary and read your financial records,
read your love letters, read everything, but before leaving the house, it imploded?”
And he said, “That’s not violating privacy.”
I immediately said, “Of course it is. Yes, it is.” And he said, “No, it isn’t.
Nothing’s kept. Nobody knows about it.” I said, “That robot has read everything.
Does that robot know if I’m sad or if I’m feeling fear, or what’s happening?”
And he looked at me and he said, “Oh, no. That robot knows a lot more than that.”
> Nobody knows about it.
This idea that no person knows about it is suspiciously similar to the word games the NSA plays with their claims of only needing a warrant once a human searches for something they already captured and stored.
Coupled with the fact that the function of said robots could be changed with a code push...
Doesn't that create an incredible opportunity for monetization and therefore risk for future abuse?
Just because we could provide a perfect search result to any query if we knew everything about a person (including their current thoughts) doesn't make enacting that a reasonable or good idea.
At the very least, Googlers :believe: they're 'doing the right thing', so they :believe: their marketing and their mottos, at least.
It's up to us to convince Google that it's wrong.
>This is not connected to any specific person
How can data not connected to any specific person be considered personal information? I mean what's bad here, the fact they collect it and then scrap the personal information? Or the fear they probably don't?
Of course not. Google still has that individual's data and can personally identify him. It's just their policy not to use that data when using machine learning and whatnot to get a result out of it.
Also, such "anonymized data" is never really anonymous, until Google starts using something like homomorphic enryption or client-side encrypted databases:
De Montjoye and colleagues examined three months of credit card transactions for 1.1 million people, all of which had been scrubbed of any PII. Still, 90% of the time he managed to identify individuals in the dataset using the date and location of just four of their transactions. By adding knowledge of the price of the transactions, he increased “reidentification” (the academic term for spotting an individual in anonymized data) to 94%. Additionally, women were easier to reidentify than men, and reidentification ability increased with income of the consumer.
https://hbr.org/2015/02/theres-no-such-thing-as-anonymous-da...
The problem is that Google gets that data in the first place and stores it - indefinitely. If someone hacks Google it could also get that data. And what's worse, instead of making it a company priority to "fix" this, it doesn't seem to care very much, even when the solutions could be within reach with a little more R&D:
Lloyd made his pitch, proposing a quantum version of Google’s search engine whereby users could make queries and receive results without Google knowing which questions were asked. The men were intrigued. But after conferring with their business manager the next day, Brin and Page informed Lloyd that his scheme went against their business plan. “They want to know everything about everybody who uses their products and services,” he joked.
So Google is like a child that wants to eat candy, has candy, but doesn't eat it?
Of course they're using all the data they can, to the extent they can.
They're just telling us there's no reason to worry, while they feed their data to the NSA's surveillance machinery etc.
You see, Google is not the only "child with candy". Then there's the US government, and that kid has a ravenous appetite.
Lowering a broken page's rank so that students only see working pages helps their educational experience.
I'm not saying I completely agree with the argument, but it isn't a bold face lie.
(i) Pages like https://fpf.org/issues/ad-tracking/ don't present a critical and balanced view on ad tracking but put forward the view of ad companies.
(ii) A random quote from the page on big data: "In Big Data: Catalyst for a Privacy Change, Joseph Jerome discusses the rise of big data and the role of privacy in both the Fourth Amendment and consumer contexts, and argues that the future of privacy will have the be built upon a foundation of trust." The future of privacy needs to be built on a foundation of trust? Well, you can state that, but it indicates that you're not an independent thinktank but try to advance policies that are beneficial to your supporters
(iii) Supporters (https://fpf.org/about/supporters/) all big shots, but also including Facebook, ComScore, ad companies and so on.
So take it with a grain of salt that the Forum for the Future of Privacy states that Google has done nothing wrong.
https://fpf.org/about/our-mission/
Their "About Us and "Mission" pages are identical, and are full of glowing talk about convening a discussion that "seeks to advance responsible data practices", and doesn't say anything about privacy. This smells strongly of a corporate/lobbyist PR front.
Some examples:
- Google Hangouts is one of the covered apps. But to be actually useful (e.g. if you want to share images), you had to turn on Google+. Images end up in Google+ photo albums, which are not covered by the Google Apps agreement and the data can be mined for advertising purposes.
- Google Drive is also covered, which means that data on Drive is not mined for ads. However, in a realistic work environment, you have to share files outside your organisation. If you share a file and someone ends up opening it on a non-Apps Google account, your document has been leaked to an ad-mined Drive account.
- When Google Photos was introduced, suddenly all my Drive photos were there as well. Google Photos is not covered, which means that the photos can be mined for ads. Luckily, this problem seems to have been resolved.
- Google Mail is covered by the terms of apps. But if I go to Google Plus, people I communicated with via e-mail are suggested as circle additions. Since Google Plus is not covered by the Apps terms, why does data from e-mail end up there?
My impression is that (perhaps driven by the consumer apps) data isolation is a real problem within Google and there is a certain amount of data shared between apps that are covered by the Google Apps terms and those that aren't. This seriously undermines my trust in Google Apps and its suitability for business/education.
Of course, I could switch off every Google service in Apps outside what is covered. But then you need two Google accounts, which is really inconvenient.
It is maddening.
For example, if you do not agree to g+ on your main account and by mistake you select that account on the dropdown from hangouts or the play store* then you get the EULA agreement and there is no way to go back to the app and select another account if you don't agree. Because as soon as you decline you exit the app, and opening the app again will how the aggrement for the last selected account, if you having another account already green lighted. The workaround? remove one account. clear data on that app, and redo everything.
* there is no way to NOT show the main (undesired) account on the drop down, and you have to use the dropdown a lot on hangouts as SMS is another account! so every time i move from sms to wifi calls, i have 1 in 2 chances of screwing everything, even more so as my profile pic on my main account is not shown because i never agreed to some policy that would allow it
i was mostly talking about android apps, where there is hardly an alternative for google services.
This is one of the main reasons why I don't use Google services, except for their search engine.
It was before the digital age too ("What do you mean your phone number isn't in the phone book?"), but people are more aware of it now.
Basically, it boils down to:
- Google is violating the student privacy pledge by logging non-educational material a student user sees, by having sync turned on by default, and by allowing administrators to turn on sync if it is off
- Student Privacy Pledge is enforcable by the FTC due to previous precedent
- EFF wants FTC to open an investigation into Google's practices, and, if found in violation, to stop the offending practices and delete already-collected personal data.
It seems fairly straghtforward to me.
I have several classes that require the use of proprietary software such as Google classroom, Microsoft silver light, and Adobe flash. On top of that I cannot even remember the last time I was given a document that wasn't a .docx.
It would be one thing if these tools were optional, but they are strictly required and students' grades do suffer if they choose to opt out of using them. In previous years I simply followed along in my windows VM, but after getting accepted to the university of my choice I've had enough. I have alerted all of my teachers I will not be using Google classroom, or complete any assignment that requires the use of flash, silver light, or any other proprietary software. Has my grade suffered? Sure, but now that I am securely in the school of my choice I could not care less.
Society is living with compromises. If you want everyone to switch to open source software make it better than the alternatives, they won't switch because of any other reason.
If you were on the wrong platform you had a hell of a time submitting assignments. (CS students on Windows found their C programs would run fine on their laptops but would crash in the CS environment because of compiler and C std lib differences. Engineering students on Linux found that they couldn't open the .docx files to submit lab reports without destroying the mandatory formatting in the tables.)
[0] By "they" I mean the persons who created the documents, not the ones who try to edit them using Gnu/Linux...
Often when people talk about open source "not being as good" what they really mean is "not being the exact same" if a button is in a different spot or something is named differently that is perceived as being "worse" because for the most part people hate change.
GAFE requires a dns change, and optionally syncing your directory with them, and has a modern interface for the admin user. So to be easier I'd probably want a debian package, or a virtual machine image.
If someone would package a virtual machine image that interfaced with common directories, provided web office tools, and a version of Classroom (which doesn't have many features), for free, it would probably have some traction, not needing to rely on the internet connection at school to be stable.
I think it wouldn't be hard to provide a virtual machine image that provides similar functionality to Google Apps (at least e-mail, storage, calendaring, and address books). Educational institutes switch to GAFE because it cuts away a lot of sysadmin salaries. You do not have to worry about getting your SMTP server whitelisted when a student/employee falls for phishing, ensure redundancy, scale up the architecture as users have more data, replace malfunctioning servers, etc.
Of course, Google's competition to an open source solution is not really fair. First of all, Google can benefit the economy of scale. Secondly, Google is large enough to assume the cost in trade for future income (get them hooked or locked in). Thirdly, Google can to use the data indefinitely to some extend for their own use (e.g. training models).
An open source solution can only win over GAFE (or perhaps Office 365) when a principled choice is made to protect the privacy of students and employees.
tl;dr: it is hard to compete with free.
However, your objection to the use of proprietary software seems completely misguided to me and reeks of overzealous naivety. The software you describe are tools for educating and conveying information. Why don't you object to your being forced to use other proprietary, non-software tools? Your teachers "force" you to use the textbook they select for the course because if every student used a different textbook, the teachers' jobs would become impossible. Why do you accept the "forced" textbook but not the forced software tool?
Your teachers force you to use a particular software tool because education would be hopelessly inefficient if every student were using different tools.
Textbooks commonly reflect the biases of their authors. These biases range from benign (Sedgewick doesn't like parent pointers, so he rants against them in his book) to malevolent (See any social studies text in Texas).
Without sufficient knowledge of the subject, a reader may be unaware of the author's biases and will shape his view of the subject based on biased information. Most high school students (and people in general) do not have sufficient knowledge of history to carefully analyze the biases of their author, nor do they have enough time to read enough books to access a balanced view.
We accept the negative effects of this bias when the effects are minor, as is the case with most high school textbooks. The benefits of forcing everybody to use the same textbook generally outweigh harm caused by bias.
Our daughter and son in law are also similarly careful about our grandkids' use of school systems and the Internet in general.
We now have the capability to algorithmically generate an estimate of how long it should take an average person to read any given text. Take something like 3/4 of that time (this value is tunable), and if a person wasn't given at least that much time to read a contract, then there patently cannot be a "meeting of the minds"[1] and the contract was never established.
This removes much of the social pressure against holding up the line, and discourages extra-long contracts. Every contract should be read and understood completely.
Good luck with your college.
For those that aren't aware, it's essentially a solution to the issue of BYOD. Students want to use their own laptops to take exams (and I'm sure it's also a lot more economical for the school since each student will already own a computer whereas building large computer labs and maintaining many machines just for computerized testing would be prohibitively expensive in our current facility).
The way it works is you essentially grant the software low level root/admin access to your machine so it can sandbox exam software and prevent you from running anything else during tests. This alone would drive me batshit insane if I was a student forced to use this setup but the worst part is that the actual software always seems out of date and doesn't support current OS versions.
We had so many problems this term when students arrived in the fall with new laptops running Windows 10 and OSX 10.11. Apparently, ExamSoft hadn't had enough time to make sure their software worked properly with these new releases so students were told a) if you haven't upgraded to the most recent OS version yet, don't do so. b) if you have recently upgraded to the most recent OS version, you should downgrade. c) if you just bought your computer with the latest OS version installed, um...we'll try to work something out.
I am so glad I am not in the group that handles this stuff because it seemed like a total shit show. The nature of the software prevents it from running in a VM so you can't just run an instance of Windows 8 or OSX 10.10. A lot of the exams aren't simple computerized versions of the old scantrons or paper tests and require the use of software run from inside SofTest. And the worst part is that people generally want to keep their OS fully updated and patched for security and support purposes (not to mention the money people spend on new computers only to find that the required software only runs on old OS versions).
From my end, dealing more with online course delivery, I've thankfully been able to move us away from Flash or Silverlight players for watching recorded content but there's still infrastructure that we need to work around since we can't simply replace it all if the frameworks in place still require some plugin or another for the time being. Not so much a privacy issue as one of keeping up with advances and standards when enterprise/education software can be slow to adapt.
Let's say this whole pledge thing didn't exist and the only thing that GAFE promise is no ads. Now let's supposed Google is collecting and anonymizing usage data, like most sofwares do, is that a big deal? Since 99% of the people won't read the ToS and most people assume some level of usage data will be collected, what would be people's stand on that?
I think, Google's stand on the current issue is that they disagree their intent violate the pledge. They would argue data collection is anonymized, and like most software out there (Chrome, FireFox, etc) these are telemetry (and all sorts) collected. For Google the collection is for regular development purpose. Perhaps they will argue that they did not do a good job at making a "disable me" option, and that they didn't make that clear to their customers. I supposed EFF just doesn't trust Google handling the data in the first place, since Google can lie about not logging the individual user name in some server's log file.
I'd be surprised if it was just "for regular development purposes."
I'd wager that collected usage information in drives marketing and sales efforts, to enterprise customers at least but perhaps consumers as well. And while it's not directly used to target ads at the students on that platform, I'd be similarly surprised if it wasn't used to train their commercial-side profiles to make targeting "anonymous" 12-18 year-olds more effective.
But if they will provide an opt-out option, that will help resolve the issue. EFF isn't wrong, they are catching an implicit fineprint, and Google is interpreting that very differently (not being Google).
Maybe they are not (currently) using the data in the very legalese-specific way that they have claimed. But, there's no doubt that they will build unrelated products using that data to make their owners gazillions. That's fine, you say - we're all here to make money, right? But, imagine that you're tracked, monitored, and experimented on 24/7 throughout your childhood. When you're 20, will you be surprised when Google's AI will understand every intention, desire of you and be able to recognize every nuance of your actions and behavior? To me, that's scary.
In the comments, a parent advises their kids to only use their Chromebooks for class activities and never for personal email and web browsing. Great advice, and should work fine for children with privacy respecting smartphones.
GAFE is really convenient in a lot of ways; the EFF is absolutely right to complain though that there are a lot of privacy issues baked into the whole thing by virtue of it being Google Services.
Outside of data collection, the issue is that GAFE admins are still at the mercy of Google changes. We may get advanced notices from Google about changes and certainly some are optional, but things like G+ integration, Photos, and so on are completely outside of our control.
My last workplace used GAFE for the University, and we had a lot of very basic privacy things that popped up as a result of GAFE and new gmail features pushed out. Our few privacy minded users (mostly older professors) would get really upset over how much data got sync'd between LDAP and G+ without our admin realizing it, students who had legitimate privacy concerns (harassment cases) were frustrated with the same issue, and so on.
I'd like to believe Google could just ease up the data collection from schools and businesses - it can't be that great compared to the public at large, and it would make GAFE absolutely perfect for schools. (Seriously, privacy aside, GAFE is otherwise perfect for a University set up. Unlimited space, great for student collaboration, and honestly many universities could do away with Office subscriptions at this point since the Google Office Suite is close enough for all intents and purposes.)
Cue endless SEO community debate & speculation in 3...2...1...