12 ms·
Browser Fingerprinting: A Survey
- harryking 8y agoWell a nice initiative i would say
- Santosh83 8y agoCan the major browser vendors not get together and implement a set of standard values and defaults that can be activated to reduce one's fingerprint, at least enough to make it non-unique? Something like what Tor Browser does but expanded to include all the major browsers. The way is open, the question is if there is will.
- givinguflac 8y agoApple has done this with safari, now we need other vendors to follow suit. I wouldn’t hold your breath given the near-monopoly chrome has as of late.
- user17843 8y agoit hasn't, it was PR. look up amiunique.org or panopticlick.eff.org with Safari.
- Wowfunhappy 8y agoYep, I was super disappointed when I tested this.
- user17843 8y agothe marketing department probably decided that it needs to be done and the safari engineers realized this would completely destroy usability for a majority of users, so they quietly abandoned it. Especially as it isn't as simple as a standardized list of fonts when the Canvas hash is 100% unique for everyone. Still very disappointing.
- jwandborg 8y agoThe engineering department should have told the marketing department that they didn't implement it.
- saagarjha 8y agoThe Safari team has not "abandoned" this effort: they are still working on reducing the fingerprinting surface area (of recent note, Safari has removed its "do not track" preference, which was used as a fingerprinting datapoint, no longer presents installed third-party fonts to websites, and no longer supports most plugins). The issue is that they have not "solved" the issue yet.
- user17843 8y agothe test shows unique system fonts for me. Here's a quote: "There will also be new security measures to prevent digital fingerprinting, or the use of things like installed fonts and plug-ins to help track users across the internet even with privacy settings active. Websites will be given a stripped down, simplified system configuration so every user's Mac looks like every other user's Mac."
- saagarjha 8y agoThis is what my modern Safari exposes: Andale Mono, Arial, Arial Black, Arial Hebrew, Arial Narrow, Arial Rounded MT Bold, Arial Unicode MS, Comic Sans MS, Courier, Courier New, Geneva, Georgia, Helvetica, Helvetica Neue, Impact, LUCIDA GRANDE, Microsoft Sans Serif, Monaco, Palatino, Tahoma, Times, Times New Roman, Trebuchet MS, Verdana, Wingdings, Wingdings 2, Wingdings 3 (via javascript) I have hundreds more fonts installed. Panopticlick says that about 1 in 14 browsers have this value, so your browser should have it too.
- user17843 8y agoI have to withdraw my original statement, Looks like I have the same set of fonts as you. So that part seems to work, I was just confused because the effect on overall uniqueness is very low. Still I applaud the efforts by the WebKit Team.
- zzzcpan 8y agoBut this is how big companies behave. They have this unwritten policy to justify any anti-competitive, anti-user behavior with "security" and "privacy" PR.
- merpnderp 8y agoHow does this work? For instance I have an iPhone model, so why would screen resolution help increase identifying bits coupled with the user agent? Everyone with my phone version and OS version will have the exact same values.
- gruez 8y ago>For instance I have an iPhone model, so why would screen resolution help increase identifying bits coupled with the user agent? Everyone with my phone version and OS version will have the exact same values. The user agent for mobile safari doesn't identify the iphone model, only that it's an iphone[1]. Knowing the precise model definitely helps to fingerprint more. [1] random search: Mozilla/5.0 (iPhone; CPU iPhone OS 12_1 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/12.0 Mobile/15E148 Safari/604.1
- benplumley 8y agoApparently one in 900 browsers has my list of system fonts, by far the most identifying thing Panopticlick could detect. However, the font list looked completely vanilla to me. I wonder if there's a telltale one in there that I installed manually.
- lmkg 8y agoAre they in alphabetical order? The order of installed fonts used to be a dominant attribute in browser fingerprinting, before some browsers took steps to normalize it. Other than that, things like presence/absence of common variants like bold, and possibly some common fonts come packaged with office software rather than the OS.
- wongarsu 8y agoYou can have what Tof browser does by using Firefox and setting privacy.resistfingerprinting in about:config. But be prepared to solve a lot of recaptchas, and don't hold your breath on Google implementing something similar in Chrome. Google is still an adtech company. They didn't spend millions (if not billions) on making and promoting Chrome only to support fingerprint resistance.
- tptacek 8y agoProbably not; that task is equivalent to coordinating between browser vendors to make their software behave in every way identically, and these are generally radically different pieces of software.
- saagarjha 8y agoIdeally every browser implements web standards the same and looks identical…
- tptacek 8y agoYou can't even get TCP/IP implemented consistently, which is why we can passively identify operating systems just by observing packets. What hope do you have of getting browsers to be indistinguishable?
- saagarjha 8y agoPractically, not much. Ideally, though…
- tptacek 8y agoThis is the problem with anti-fingerprinting. It's a genuinely hard problem, not in the "it takes a lot of code and a lot of will to get it done" sense, but in an "it's unclear what computer science has to say about this problem right now" sense. It's up there with DRM/antimalware (two sides of the same coin).
- zzzcpan 8y agoThis will only shift efforts to extract uniquely identifiable information through side channels instead of directly. So you also need to disable javascript or redesign it to prevent side channel attacks (it's possible, but would be a completely new javascript compiler with not entirely compatible language, also this would mitigate spectre attacks). However first you need some serious regulations to split Google and take away its browser "business". As it is Google is doing everything to preserve and advance tracking and fingerprinting in web browsers.
- user17843 8y agoIt looks like the prevalence is low and it can easily be blocked with blocking the scripts in question. Nevertheless, what needs to happen is that all major browser makers come together and simply create a set of standard API values that do not harm daily browsing and make it possible for users to blend in with the masses, if they opt-in to activate It would be sufficient to create a couple of uniform user agents, list of fonts, list of plugins, canvas hash, platform and webgl data to bring the uniqueness down.
- amelius 8y agoI suspect it will always be possible to detect at least the brand of the browser, regardless of what the user-agent string says. It's just too damn hard to implement the entire web specification consistently, and malevolent actors can easily exploit any small deviations in the implementation. And of course things which are not in the web specification (such as timing) can be exploited.
- wongarsu 8y agoAfter you figured out the browser, it will also always be possible to find out the browser version, because newer versions implement features not present in older versions. Then you just need a hint to the OS (at least distuingishing Linux from Windows seems easy) and you have reconstructed the entire User Agent string.
- amelius 8y ago> it will also always be possible to find out the browser version, because newer versions implement features not present in older versions True, but most browsers have an auto-update scheme, so basically everybody is on the same version at a given time (by approximation).
- viseztrance 8y agoThere are SaSS companies that have their entire business model around creating databases of fingerprints - for example determining bots. These scripts track you over multiple websites and are very well concealed, so you won't find them on any ad blocker blacklist. Also, living in the EU won't protect you from getting fingerprinted.
- wongarsu 8y agoAccording to the paper the most useful attributes for browser fingerprinting are: - User agent - List of Plugins - Screen resolution and color depth - Canvas - WebGL Renderer (~GPU Model) Maybe Canvas can be fixed to a degree, but the others feel like things that will inherently be different. Firefox with the privacy.resistfingerprinting setting solves all of these at least partially (and more), but not without some usability tradeoffs.
- thejohnconway 8y agoSurely user agents could give out way less information (virtually none in fact?).
- Fnoord 8y agoYes, but it will break functionality because servers use it to apply hacks on the website before the content is served. WebRTC is another one which you can disable. JavaScript, in general, as well, requiring uMatrix. Chrome leaks the extensions you use, Firefox does not (AFAIK). As for blocking canvas, see [1] To test your fingerprinting I can recommend IPleak [2] The problem of this all is, its a usability nightmare. [1] https://addons.mozilla.org/en-US/android/addon/canvasblocker/ https://addons.mozilla.org/en-US/android/addon/canvasblocker... [2] https://www.ipleak.net https://www.ipleak.net
- Jonnax 8y agoYes it really should. Chrome on Android reports your phone model and build number. If you're using a lessor known phone with a carrier specific ROM, then you're in a really tiny population.
- titzer 8y agoI've given up on technological solutions to this. We need a "do not track" bit for browsers and it needs to carry the weight of law behind it. Violators need to be named, shamed, and fined.
- paulryanrogers 8y agoWith robust enforcement. Otherwise it's a technical cat and mouse game, and few large players are on the users' side. EDIT: typo
- gregw134 8y agoIronically, the do not track setting itself can be used in browser fingerprinting.
- 55555 8y agoIn fact, it is _primarily_ used as a point of entropy for tracking, and does almost nothing to prevent tracking, so all-in-all it's a huge net negative. And because it's mostly turned off, turning it on to be tracked less actually causes you to be tracked more.
- titzer 8y agoExactly, this is why it needs the force of law behind it.
- amelius 8y agoCan't we give penalties to companies that use fingerprinting scripts (or include them by third parties)? I'd like to see an organization that checks for this, and gives fines, and perhaps even withdraws the right to use a domain name. And I'd like to see more responsibility with site owners for using third party code. By the way, I think a withdrawal of the right to use a name (brand) is a very appropriate way to penalize a serious privacy violations. Brands are all about customer trust, and if that trust is violated, then it seems to me only fair that the right to use a name is taken away.
- user17843 8y agoOne of the referenced studies (1-million study withOpenWPM, the most recent from 2016) notes that the usual fingerprinting scripts have basically disappeared since a public outcry and media attention following a lawsuit. Prevalence: - 1.4% for canvasfingerprinting - 0.325% for canvasfont probing - 0.0715% for WebRTC - 0.0067% forAudioContext The above were found only on the most shady of all websites, and good content blockers block all those scripts. My bet is GDPR was the death blow for this kind of scripts. For small companies without a room full of lawyers data has become a liability. So fingerprinting is now basically in the hands of google, amazon, etc.
- amelius 8y ago> My bet is GDPR was the death blow for this kind of scripts. Curious, was the lawsuit in that study also based on GDPR? So far I haven't seen much coverage of GDPR based lawsuits in the media. Also, I'd like to know if there have been studies that show that EU residents can now (after GDPR) browse the web without leaving a trail of information.
- user17843 8y agoNo. Quote about canvas fingerprinting: "Comparing our results with a 2014 study [1], we find three important trends. First, the most prominent trackers have by-and-large stopped using it, suggesting that the public backlash following that study was effective. Second, the overall number of domains employing it has increased considerably, indicating that knowledge of the technique has spread and that more obscure trackers are less concerned about public perception. As the technique evolves, the images used have increased in variety and complexity, as we detail in Figure 12 in the Appendix. Third, the use has shifted from behavioral tracking to fraud detection, in line with thead industry’s self-regulatory norm regarding acceptable uses of fingerprinting." [1] G. Acar,C. Eubank, S. Englehardt, M. Juarez, A. Narayanan,and C. Diaz. The web never forgets: Persistent trackingmechanisms in the wild. InProceedings of CCS, 2014. Other references: [10] W. Davis. KISSmetrics Finalizes Supercookies Settlement.http://www.mediapost.com/publications/article/191409/kissmetrics-finalizes-supercookies-settlement.html,2013 http://www.mediapost.com/publications/article/191409/kissmet.... [Online; accessed 12-May-2014]. [15] Federal Trade Commission. Google will pay $22.5 millionto settle FTC charges it misrepresented privacy assurancesto users of Apple’s Safari internet browser. https://www.ftc.gov/news-events/press-releases/2012/08/google-will-pay-225-million-settle-ftc-charges-it-misrepresented https://www.ftc.gov/news-events/press-releases/2012/08/googl..., 2012
- FabHK 8y agoIs anyone aware of fingerprinting mitigation (extensions?) for Safari on macOS? I'd rather not abandon Safari.
- paulus_magnus2 8y agohttps://en.wikipedia.org/wiki/Stalking https://en.wikipedia.org/wiki/Stalking Stalking is unwanted and/or repeated surveillance by an individual or group towards another person.[1] Stalking behaviors are interrelated to harassment and intimidation and may include following the victim in person or monitoring them. The term stalking is used with some differing definitions in psychiatry and psychology, as well as in some legal jurisdictions as a term for a criminal offense. This should also apply in the digital world. Doesn't it?
- codetrotter 8y agoWell, there is something called cyberstalking. https://en.wikipedia.org/wiki/Cyberstalking https://en.wikipedia.org/wiki/Cyberstalking I don’t think you will be able to convince a court that browser fingerprinting is cyberstalking though, but you could certainly try.
- z1r011 8y agoCynical mode: The law suit would be very expensive, but imho. possible to get to the next round. This elevates for several instances, until you're out of money and the lawyers can burry it.
- munk-a 8y agoAnd, if it came to a lawsuit, you'd be up against Amazon, Facebook and Google all at once (among others).
- dredmorbius 8y agoYou can choose your battles. There's no need or obligation to take on all targets, and the legal strategy of first pursuing a tractable, or even sympathetic (to your interests) target is well established. The cyberstalking industry has many players. Most are small. A target-rich environment. Shifting the investment or financing calculus, via increased investor risk, possibly including legal liability, would be huge.
- overcast 8y agoOut of curiosity, how are you guys handling anonymous user actions without fingerprinting? For example if you want to setup a user poll, without requiring a user account? Storing it in a cookie can just be deleted, which the agent can just clear and resubmit.
- lisper 8y agoBrowser fingerprints can be spoofed too. Sock puppet accounts are equally easy to create. There is no easy way to insure that people don't cheat on on-line polls. This is why (for example) on-line voting in IRL elections is a Really Bad Idea.
- sdoering 8y agoThere are multiple ways. All can be cheated. * Cookies: Can be deleted * IP-Adress: Not unique, because ISPs rotate them; also VPN * Login: Well create a second one * Methods from Universities using nth letter of name and nth digit of birthdate: Just make up a new name. Sorry - but unless you are using an analog medium or asking the questions in person the numbers can be inflated and there is no way to have 100% data quality. But in most contexts this is ok. So I would probably go the most easy way: Cookie. I would - at least not in the European Union go with fingerprinting and such stuff, as I am not sure how this plays out regarding GDPR as this would be PII you are storing.
- TheAceOfHearts 8y agoFingerprinting is not an effective deterrent either, as you can just open a different browser. The user could also have multiple devices with internet access. You could use IP address, although that only works if the user isn't on a public / shared network. It's also easily bypassed by spinning up a VM on a cloud service provider and using an SSH tunnel. Since you used polls as an example: StrawPoll.me [0] is an online poll site which lets you select different duplication checks based on your requirements. The choices are: IP, browser cookie, none, or require user sign in. They also give you an option to add a CAPTCHA. [0] https://www.strawpoll.me/ https://www.strawpoll.me/
- OrgNet 8y ago
- bfirsh 8y agoIf you’re on a phone, here’s a web version: https://www.arxiv-vanity.com/papers/1905.01051/ https://www.arxiv-vanity.com/papers/1905.01051/ (now with linked citations!)
- badrabbit 8y agoCan anyone with a legal background explain to me why I can't file a restraining order that prevents companies from stalking me online? Primarily for companies that develop shadow profiles of users but also companies like google and cloudflare where their tracking is a result of an opt-in by the site operator. Would it be difficult to prove I am in fear (rational) as a result of their stalking?
- Erik816 8y agoI don't believe you can get a restraining order against a company (just an actual person). You would ask for injunctive relief, which is basically the court telling the company not to do something, in this case, tracking you online. I'm not aware of the law in this area. I'm a lawyer, and I could at least see a plausible argument, but the only way to know for sure would be to try to sue them, or find other cases where someone has successfully done so.