Is Google Chrome Listening to You?

There have been allegations that the popular web browser, Google Chrome, can potentially listen to both your media and you, leading some to develop conspiracy theories to attract more supporters. It’s crucial to examine what Google Chrome is doing in the background and determine if it is monitoring your activities.

Since this issue has implications for everyone’s safety and privacy, it’s important to take a closer look.

Google Chrome and the Chromium Project

Chromium is an open-source web browser platform project developed by Google. A team of developers working at Google developed Chromium under the BSD license. The fundamental objective of the project is to create a fast and reliable web browser. Google uses the open-source software developed under this project, as its license permits, and adds its proprietary components to create the Google Chrome web browser.

The Chromium project is not only used by Google Chrome, but also by numerous other applications and libraries such as Yandex Browser, Qt Web Engine, and more. Additionally, various Linux distributions such as Debian and Ubuntu have a web browser package called Chromium, which is similar to the Chromium project but different from Google Chrome.

Google Chrome is not free software. But by using the Chromium application that comes with Linux distributions instead of Google Chrome, you can avoid the proprietary components added by Google.

How Was the Problem Detected?

A mobile device using Google Chrome

On May 26, 2015, Yoshino Yoshihito drafted a bug report on the Debian Bug Tracking System. In this bug report, he summarized that after updating the Chromium application to version 43, the system downloaded an extension called “Chrome Hotword Shared Module” without Yoshihito’s consent, and when Yoshihito entered chrome://voicesearch in the browser’s address bar to obtain additional information about the extension, it indicated that it had “Audio Capture Allowed” permission and therefore had access to the microphone.

After the publication of this bug report, many people shared information confirming the same situation. So what does all this mean, and where does the problem start for us?

What Do We Know About the Google Voice Search Technology?

Most likely, you have tried Google’s voice search feature on your computer or mobile phone. This feature starts working when you click a button that indicates you want to perform a voice search or when you open a special application designed for this purpose on your phone.

Google’s servers record and analyze your voice from the microphone and compare it with previously stored patterns. Currently, running advanced speech recognition functions directly on a personal computer or mobile phone is not practical. Therefore, it is inevitable to use cloud-based server clusters for analysis.

Google uses artificial neural network models, which are inspired by the information processing method of the human brain, to better perceive the voice pattern recorded in its speech recognition technology.

women speaking on phone

Google utilizes the advantage of its enormous data, which includes previously analyzed web content and frequently searched queries, to improve the speech recognition process with the N-gram model. The N-gram model was developed for cases where the probability of the Nth word being known is based on the previous (N-1) words. For example, if there is a third word after two words such as “afternoon tea,” the probability of this word being “mathematics” or “breakfast” differs. To increase the success rate in predicting the third word in the example, we need data to feed the 3-gram model.

According to the figures provided by Google, in 2012, it was using a 5-gram model (i.e. up to five consecutive words) and a total of around 12 billion N-gram sets for speech recognition technology. While I cannot be certain about the level these numbers have reached today, I can assume that they have increased significantly. In addition, it is worth noting that this is not a static model and continues to learn.

This detail is necessary to understand why the recorded voice is sent to Google servers. The algorithms that continue to learn on the aforementioned large datasets can produce quick responses, but this is only possible by running on multiple servers. Even if your personal computers reach the required processing power someday, Google will always have an advantage because you don’t have the necessary large databases for analysis.

Is Voice Searching Really Malicious?

Chrome browser on a samsung tablet showing the Google homepage

Voice search has been a Google product for a long time, but with Google Chrome version 43 and beyond, this process has been taken a step further and supported by a new model called OK Google. This feature works when you enter the Google homepage of your browser or open a new tab in Google Chrome to go to an address. All you need to do is say the words you want to search after saying OK Google. So instead of clicking a special button to do a voice search, you activate the speech recognition process by pronouncing the words OK Google. You can ask several interesting and helpful questions using OK Google, which ultimately makes your life easier.

For this feature to work, your system needs to constantly listen to the environment sounds and capture the OK Google keyword. The analysis here is done on your computer because the word count is only two and it contains a specific sound pattern, and no data is sent to Google servers. But the next analysis requires servers. Therefore, the way to enable you to search faster is to constantly listen to you. Moreover, you don’t receive a notification that your microphone is in use during this listening process.

With the release of Google Chrome version 46, the OK Google feature is now gone. So, has another technology replaced it, or if it has, how harmful is this technology for us, can we trust Google?

Even if Google is said to be one of the most secure technology companies in the world, if malicious hackers find a strong weakness in such technology, they can exploit it. As a result, the way such a technology works may be that environment listening is activated for certain people at desired times and continues to work normally for the rest of the time.

Also, Google’s privacy policy states that Google uses users’ voices to develop voice recognition technologies. Of course, only Google may know these voices. Or Google can even use it to catch criminals with the police. However, it is a situation that normal civilian users should be aware of.

So What Can We Do?

Google is a really powerful company that has managed to steer technology. It has almost all the web traffic in the world too. This makes it a superpower. Google is one of the investors of many voice analytics and voice search algorithms that have evolved over the years. That’s why it’s important to have doubts about whether Google Chrome is listening to you.

Because behind such large companies, there are serious financial investments, government support, and publicly traded shares. What you need to do here is to protect yourself instead of waiting for such big companies to protect you. That’s why you can always refer to open-source free software.

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button