In the age of smartphones, if you need to perform a task, there is probably an app already developed for it. From online chatting and navigating to playing games or keeping your dinner reservations neatly organized, there are countless apps out there for anything remotely related to our daily lives.
In contrast to the more selective Apple App Store, Google has to deal with a gift and a curse: most of the world’s smartphones and mobile devices run on Android. This means that there is a much higher number of mobile apps developed and launched on the Google app store, leading to a lot of junk clutter. So how does Google decide which apps no longer deserve a coveted spot on the Play store?
Machine Learning Has Been a Game – Changer
Google has, in the past, proven that user safety and convenience is taken very seriously, so it was only a matter of time before a crackdown on undesirable apps took place. In early 2017, the tech giant had announced its intention to move forward with evaluating apps on its Play Store and reduce the number of “bad apps” on offer in order to provide a better user experience – and protect its clients.
In early 2018, the company released the first results: in 2017 alone, more than 700,000 apps and over 100,000 developers were removed from its app store. These figures mark a 70% rise from the previous year – while there were claims that roughly 99% of malicious apps were effectively rooted out before anyone could install them. Android users were more than happy to hear the news, and it seems that Google will continue with this approach.
But how does the company identify which apps must go? After all, the Play Store hosts a particularly high number of mobile applications – not to mention that new apps are being added literally every day. Google representatives have stated that the industry leader has turned to the latest cutting-edge tech solution to eliminate “bad apps”: machine learning. This allows Google to automate its analysis to a large extent, taking on big data and offering pretty accurate evaluations.
By employing algorithms and machine learning, the company was able to comb through an incredibly high volume of information and pinpoint the troublemakers: apps or developers that were then removed from the app store. Yet, in order for machine learning to work, Google had to decide on identifiers – features and standards that rendered a specific app undesirable.
Google Targets Illegal Activites, Malware, Impersonation and Inappropriate Content
As any major company does, Google has to conform with specific requirements laid down by legislation across different countries, as well as comply with regulatory rules and guidelines. One of the first aspects that the company looks at is that apps on its store are in line with the rules that developers must adhere to. One of the most recent actions that Google took in this regard was the removal of two forex trading applications, as reported by RussianInsight.com on 5 February 2019.
The trading apps that were used by Russian users were operated by trading companies that had failed to receive proper authorization from Russian authorities, as requested by the country’s legislation. The Russian Central Bank asked Google to intervene, which the company swiftly did by identifying and deleting the two apps. When it comes to apps like forex trading platforms, which process financial data and money, it is extremely important for users to be properly informed before choosing a provider – as this crackdown demonstrates.
As mentioned, keeping users safe is of paramount importance for Google and this philosophy guides its decision-making process with regard to apps. Applications tied to malware are a top target in Google’s crusade against abusive software on the Play Store. These are also known as PHAs (short for Potentially Harmful Applications), which includes apps linked to Trojans, phishing and online fraud.
Even though the company believes that the percentage of PHAs on its app store is relatively small, it takes the matter very seriously and has made it one of the top priorities in fending off malicious developers. The results are already impressive, as PHA numbers have gone down dramatically since 2017. The tech firm has even launched a new security feature, Google Play Protect, dedicated to scanning malware on the Google app store.
Malware-related apps are not the only type of malicious software that Google has to tackle in its efforts to beef up security on the Play Store. The most common red flag that leads to removal of an app from the store is an impersonation. Many abusive apps attract their victims by mimicking famous apps and make it seem like they are the real deal, or somehow connected to the authentic title. Famous mobile applications get a lot of traction and traffic by users who search for particular keywords tied to them so copycat apps try to take advantage of that in order to trick users into installing them.
Finally, inappropriate content is definitely a no-go on the Play Store. Any apps that violate Google’s terms for what type of content is not allowed, like depicting extreme violence or illegal activities, are automatically flagged by the machine learning algorithms. Human reviewers then take over to assess the situation and detect problematic apps that should be banned.
Enforcing the company’s rules for mobile applications is no easy feat. But, thanks to the increased technical capabilities offered by machine learning, the task is much more manageable than it used to be – and it shows on Google’s success rate.