Brave takes on the creepy websites that override your privacy settings

Privacy antidote —

Even if you block 3rd-party cookies, bounce tracking can set them anyway. Until now.


Stylized illustration of a fingerprint.

Getty Images

Some websites just can’t take “no” for an answer. Instead of respecting visitors’ choice to block third-party cookies—the identifiers that track browsing activity as a user moves from site to site—they find sneaky ways to bypass those settings. Now, makers of the Brave browser are taking action.

Earlier this week, Brave Nightly—the testing and development version of the browser—rolled out a feature that’s designed to prevent what’s known as bounce tracking. The new feature, known as unlinkable bouncing, will roll out for general release in Brave version 1.37 slated for March 29.

Overriding privacy

Bounce tracking is one of the key ways websites circumvent third-party cookie blocking. When a browser prevents a website such as site.example from loading a third-party tracking cookie from a domain such as tracker.example, site.example pulls a fast one. When site.example detects that the tracker.example cookie can’t be set, it instead redirects the browser to the tracker.example site, sets a cookie from that domain, and then redirects back to the original page or a new destination.

With that, the tracker.example cookie gets passed through a URL parameter and then gets stashed as a first-party cookie on the landing page. Once tracker.example places itself between enough of the sites a visitor browses, the tracker eventually builds a detailed profile of that activity, including the user’s interests and demographics.

The image below shows how third-party cookie blocking is supposed to work. When the user moves from site-one.example to cats.example and later from site-two.example to cars.example, there’s no way to track those movements as coming from the same person.

Bounce tracking circumvents this arrangement by inserting a third-party tracking site such as tracker.example in between the originating site and the cats.example or cars.example sites the user later browses to. Tracker.example then records that it was the user who visited both cats.example and cars.example.

While browsers that support third-party cookie blocking have existing mechanisms designed to thwart bounce tracking, this sneaky form of surveillance remains hard to defend against, since the browser doesn’t know beforehand that it will be directed to tracker.example. That’s where unlinkable bouncing comes in.

Ephemeral storage to the rescue

In a post, the Brave privacy team on Wednesday outlined the process that unlinkable bouncing uses. In a nutshell, unlinkable bouncing checks the site a user is about to visit against a list of URLs known to perform bounce tracking. When a destination site appears on the list and Brave has no cookies, localStorage, or other data related to it, the browser automatically creates a new one-time browser storage area for the site.

Once a user leaves the tracking site, Brave deletes the temporary storage. Because the data is no longer stored, the tracking site will be unable to re-identify the user the next time they are bounced through it.

Brave has several other ways to prevent site tracking. They include query-parameter stripping, debouncing, and (when blocking is set to aggressive mode) a warning to give concerned users a chance to back out.

The Brave privacy team explained the full flow as follows:

  1. When navigating to a new URL, Brave checks to see if that URL is a known bounce-tracking (or otherwise harmful) site, by consulting filter lists (both crowdsourced and Brave-generated).
  2. If that URL appears in a filter list, the browser checks the Trackers & ads blocked shields setting for the destination site. If that setting is Aggressive, the user is presented with a warning for whether they want to continue with the navigation, as described in a prior blog post.
  3. If the user has Trackers & ads blocked in the default setting (or decides to continue with the navigation in the Aggressive setting), the browser then checks the first-party DOM storage values (cookies, localStorage, etc.) for the destination site. If the user has any existing stored values, the navigation continues using the existing stored values (in other words, Unlinkable Bouncing is not applied). If no DOM storage values exist for the destination site, the browser creates a new, temporary browser storage area for the destination site.
  4. Soon after you leave the suspected bounce-tracking site (meaning no tabs are open for that site) the temporary storage is deleted, preventing the site from re-identifying you the next time you’re bounced through the site.

Team members said that unlinkable bouncing is the first of four planned applications to implement what they call “first-party ephemeral storage.” The set of techniques allows a site to identify visitors for only as long as they have it open. As a result, first-party ephemeral storage prevents the first-party site from re-identifying a user unless the user wants to be re-identified.

Using first-party ephemeral storage will be akin to clearing browser storage every time the user leaves the site, except it’s easier and more targeted.

“This brings about a total shift in the Web’s default behavior,” the privacy team members wrote. “To date, browsers have assumed users want every site to remember them unless the user takes some explicit step against that remembering. Instead, Brave is working toward forgetfulness (and thus privacy) by default.”

Note: This article have been indexed to our site. We do not claim legitimacy, ownership or copyright of any of the content above. To see the article at original source Click Here

Related Posts
AI 走入微型裝置 掌握機器學習 thumbnail

AI 走入微型裝置 掌握機器學習

不少學校開始投入教授 AI 方案, 而獻主會聖母書院早於兩年前已投入於此,這次於暑假期間向筆者分享最新方案,利用 Arduino 將 AI 方案收納其中,務求讓學生於學習時感覺新穎,而設備也有毋需連線及粍電量低的好處, 務求可以較低的門檻讓學生接觸機器學習 (Machine Learning) 。 早於 2019 年,獻主會聖母書院已有發展 AI 教學方案,起初是運用 Raspberry Pi 4 加 Pi Zero Camera ,配合 TensorFlow Lite 製作手語方案,令學生具備基本 AI 認知。次年因應疫情,製作社交距離檢測儀,運用雲端的 Google Colab ,學習物件辨識的運用和盲點。該校過往的教學設計,大多採用 Raspberry Pi ,優點是價錢合理,每件約 400 多港元,但往往要配合特定的作業系統,對新手而言,學習門檻較高。 獻主會聖母書院教師團隊(左起)教師鄭臻諺、彭嘉煒和黃梓駿。傳統程式與機器學習的分別。機器學習降落手中 至於該校新發展,彭嘉煒老師講解,這次會讓學生理解傳統程式與機器學習的分別,傳統程式如要辨識貓,就要命令電腦按規則分析,如顏色、形狀等。機器學習 (Machine Learning) 角色不一樣,把答案和樣本給電腦 「學習」,電腦自行找出規則,如貓有甚麼特徵?顏色是怎樣?形狀是怎樣?經過機器學習後,由電腦決定甚麼規則才有效。至於這次的學習課程及計劃籌備,由黃梓駿老師負責,他運用 TinyML (微型機器學習技術)作全新教學方案。黃老師表示,隨著人工智能成熟,人工智能在雲端以外,逐漸走入至微型設備中,如此一來不用經由雲端資料傳送,加上在電量方面著重低功耗發展,有望在更多設備上應用。此外, TinyML 能儲存於 Arduino ,成本更低之餘,操作難度亦降低。 黃老師選用 Arduino Nano 33 BLE…
Read More
Herní lov za 85 miliard dolarů. Po obřích nákupech Microsoftu, Sony a tvůrců GTA visí otazník nad malými studii thumbnail

Herní lov za 85 miliard dolarů. Po obřích nákupech Microsoftu, Sony a tvůrců GTA visí otazník nad malými studii

V novém roce uplynulo sotva čtyřiatřicet dní, ale už se toho na globálním herním poli stalo tolik, že by to vystačilo i na celých dvanáct měsíců. Jen během ledna se odehrály tři velké obchody, které mají potenciál proměnit celý segment. Lépe řečeno jej zkonsolidovat, což je situace, která se ve spojení se společnostmi Take-Two, Microsoft…
Read More
Motorola Moto G Stylus (2022) thumbnail

Motorola Moto G Stylus (2022)

Wygląda na to, że znana i lubiana seria smartfonów Galaxy Note od Samsunga zakończyła już swój żywot. W zamian koreański producent przerzuci rysik do czołowego modelu z linii Galaxy S, jednak poniekąd oznacza to, że taniego urządzenia w tym stylu już raczej nie znajdziemy. No, chyba, że spojrzymy na to, co prezentują inni producenci. Od…
Read More
Vivo V27 Pro May Launch in India on February 25 thumbnail

Vivo V27 Pro May Launch in India on February 25

        | Updated: Friday, February 17, 2023, 14:02 [IST] Vivo is readying its next big launch in the mid-range segment in India. The company confirmed through a social media campaign on Thursday that it will unveil the Vivo V27 series soon. A few days ago, Flipkart also added a microsite for the
Read More
Study: Inflammation drives social media use thumbnail

Study: Inflammation drives social media use

BUFFALO, N.Y. – Inflammation is the body’s response to injury and infection, but it is also a factor that can lead people to use social media, according to a new study by a University at Buffalo communication researcher.Across three studies involving more than 1,800 participants, the findings — published in the journal Brain, Behavior and
Read More
Index Of News
Total
0
Share