Google Chrome Taps Machine Learning to Curb Phishing

Credit: BiljaST

Google has been leveraging machine learning to enhance its products. Its recent venture is to boost the security of its browser Google Chrome. The company rolled out a new machine-learning-powered model that it said distinguishes malicious websites and phishing attacks 2.5 times more than the previous one. As a result, the Google Safe Browsing service should ensure safer browsing. Besides this, Google is also planning to silence pesky web notifications based on how users previously interacted with them. This further improves security since some notifications carry malware.

When using Google Safe Browsing, Chrome users receive a warning should they attempt to access phishing or compromised sites. This bars access from the said portals unless users adjust the settings and switch off Safe Browsing. Disabling the service can be risky considering that the average cost of a data breach is a hefty $4.24 million.

To assure the safety of its users, Google maintains a transparency report that displays the number of warnings sent per week from 2010 to May 2022. It also details the website owner response times from 2006 to February 2022. Included in the report is a diagnostic tool with which users can check if a website is safe to access. This comes as a preview of Safe Browsing’s proficiency since it’s powered by the same machine learning algorithm.

The release of Safe Browsing is a big deal for internet users. After all, as of May 2022, Google Chrome accounts for a staggering 64.95% share of the browser market, with its closest competitor, Safari, making up only 19.01%. Google also notes that it offers protection to more than 4 billion devices. More importantly, users are set to receive protection from one of the most dangerous digital threats around, phishing.

Understanding the Danger of Phishing Attacks

Phishing is defined as a practice of fooling internet users into disclosing confidential information, which can be used illicitly. They can come in the form of deceptive emails and websites that ask for or scan computers to obtain information like credit card numbers, banking details, and the like. As such, the coverage of Safe Browsing includes Chrome, Gmail, Android, and other browsers. These comprise the areas that are typically vulnerable to phishing attacks.

As seen in the Google transparency report, the number of daily warnings sent is staggering, which trails the actual volume of attacks. Recent cybersecurity statistics reveal that an estimated 3.4 billion phishing emails occur are sent each day. In addition, 214,345 unique phishing sites were identified in 2021. What’s alarming is that, in 2020, the FBI discovered that there were 241,342 phishing victims in the US alone. Moreover, another study found that 33% of phishing emails are opened. Phishing is also the most common type of cyberattack experienced by US companies.

Based on the data, the ones who need Google Safe Browsing the most are SaaS companies, financial organizations, and social media users. This is reflected by a 2020 APWG study that reveals SaaS companies, at 34.7%, are the most frequently targeted organizations, followed by financial institutions (18%), payment gateways (11.8%), social networks (10.8%), and e-commerce sites (7.5%).

Safe Browsing won’t likely detect every fraudulent site or email encountered. But it affords users sufficient protection from basic threats. After all, according to Google, its machine learning technology contributes to a 99.9% accuracy rating in spam detection, on top of the enhanced identification of phishing sites.