Today, we’re announcing a multi-year initiative to build the Privacy Sandbox on Android, with the goal of introducing new, more private advertising solutions. Specifically, these solutions will limit sharing of user data with third parties and operate without cross-app identifiers, including advertising ID. We’re also exploring technologies that reduce the potential for covert data collection, including safer ways for apps to integrate with advertising SDKs.
The Privacy Sandbox on Android builds on our existing efforts on the web, providing a clear path forward to improve user privacy without putting access to free content and services at risk.
A plan for a plan that aims to please the advertising industry, an industry which at this point means Google, all built by Google.
I might be mildly skeptical.
I understand the skepticism, and I think its well warranted. It seems like a really interesting engineering problem so I think most of the information you’ll see right now was written by well meaning engineering folk. For instance I think this is pretty interesting: https://developer.chrome.com/docs/privacy-sandbox/floc/
However, this just means that the bean counters haven’t gotten involved yet. It seems too good to last. Google still primarily is an add company.
Bill Shooter of Bul,
I don’t follow this too regularly as I’ve already switched to alternative platforms, but I do get suspicious of privacy initiatives by google/microsoft/apple/facebook/etc when consumer interests are in direct conflict with corporate ones. I worry about double standards where they create one set of rules for themselves and another for everyone else. It’s all too easy for them to grant themselves data mining backdoors or privileged access. This seems justified given their history of finding “creative” ways to track users without their knowledge.
https://www.zdnet.com/article/google-admits-tracking-users-location-even-when-setting-disabled/
https://www.zdnet.com/article/google-well-track-your-offline-credit-card-use-to-show-that-online-ads-work/
Are things different this time? Trust is hard to earn and easily lost. Frankly I don’t trust PR pieces from any of the tech giants right now. They say what users want to hear, but they’ve all been deceptive to various degrees and hide behind vague and unintelligible terms.
What I have seen is, more data does not necessarily mean better information.
Yes, ML engineers just want to throw everything at the model, and hope for the best. But being smart about it usually pays better.
For ads, you might have full user location history. It might have granular data over two years. But knowing the user “generally likes to take vacations abroad” is much more valuable. And of course less noisy, easier to process, has less complexity, not to mention much better for overall privacy.
So, I don’t think Google’s interests necessarily conflict with a privacy focused vision.
(Note: These are my personal observations)
sukru,
I would say that asserting there is no conflict is being very optimistic. Corporations will always have knobs they can turn to increase profits at the expense of our privacy. IMHO the main question is whether they’re willing and able to avoid the temptations particularly if nobody’s holding them accountable.
Alfman,
Yes, it is good to have some healthy skepticism. It took years to “steer the ship” to a privacy sensitive direction, it will be hard to turn it back, but unfortunately nothing is impossible. And we have no guarantees what could happen in the next decade or two.
As for who can hold corporations accountable: I would say “you, me, customers, engineers, media, general tech savvy public”. At least that is how I personally see it.
Actually in the case of Google there is more than that. There are automated systems that keep track of and restrict access to user data. For example,
which binaries can access which rows and columns in the databases, and who can run them: https://cloud.google.com/docs/security/binary-authorization-for-borg
or making sure ML models don’t “memorize” training data, and learn more than they are normally supposed to do:
https://medium.com/tensorflow/introducing-tensorflow-privacy-learning-with-differential-privacy-for-training-data-b143c5e801b6
and many more…
I think that is all I can tell…
sukru,
Yes, you could say that about every industry, but there are limits especially when it comes to duopolies, etc. We need healthy competition to make the consequences of displeasing users carry a real risk of those users leaving. In the mobile market the competition has practically collapsed to a binary choice and consumers who aren’t happy with either of the companies at the top don’t have great options.
It’s true, there is plenty of reasearch about how privacy respecting algorithms can work. However as an outsider it is a problem that we don’t know the extent of the machinery and we don’t have public auditors overseeing what’s going on. We usually don’t know when our data is used, by who, nor how. Regardless of the fact there are ways to keep private data secure, it’s hard to inspire much confidence as long as it’s being enforced by the very same companies with a conflict of interest to use it. Say google wants to buy up customer data from various sources and needs deeper access to private data to map it. Realistically they will grant themselves access to the private data if they think it will improve their bottom line and as usual users won’t be given a say.
No problem, I don’t expect you to give away trade secrets. I’m just glad that we are able to discuss these topics respectfully even though we’re coming at things from different angles.
Agreed. We, all of us, must be the ones to hold these people accountable. Remember folks, corporations have no will of their own and are run by self-interested people.
Sadly, so many people seem willing to abdicate their responsibilities and say that the government (also a group of self-interested people by the way) should do something about it instead. And then they wonder why things don’t ever change when the two groups of self-interested people figure out that they best serve themselves by serving the other group in a feedback loop which eventually runs so rampant that it’s hard to tell where the businesses end and government begins.
There is FB and the other
Data MiningSocial Media companies, but I’m not sure this isn’t a conflict of interest. It’s possible this could be construed as anti-competitive since Google owns Android and Chrome.1: Google blocks other Ad Tech companies spinning it as pro-privacy.
2: Google allows exceptions for themselves.
3: Google profits because they cut off Android/Chrome data to competitors.
I think there is a large part of this that is aimed at avoiding more fines from the EU, in the docs google repeatedly points out how they are trying to avoid reliance on them as the advertising company. That its designed to be open. Basically the browser steps in and the advertisements in the page ask the browser what the users interests are and then decides what to show. So google is setting themselves up to be that magic that lives and runs in the browser that decides what kinds of interests to tell the advertiser about. And those interests are supposedly not based directly on the browsing habbits of the user, but there is another level of indirection where the browsing habits are used to choose a profile and the profile has interests.
For me personally that would suck. I think I see a lot of that as is. Google and other ad companies see me as some one interested in tech ( duh, bingo), and then thinks, hmm I bet this person is also interested in science fiction, Anime and superhero stuff!) NOPE!!! Wrong on those.
LOL. Google: “B*** better have my money. Do they think I’m doing this for free? Nah. I’m going to get my money, or they’re going hear what my glock has to say.”