Artificial intelligence detects threats 60 times faster than humans

In the wake of a pandemic, algorithms will increasingly secure private computers as well.

The most effective cyber security solutions are supported by artificial intelligence. Thanks to algorithms, they are able to detect threats in just seconds – dozens of times faster than a human would. The human factor is also considered the biggest vulnerability exploited by hackers. The problem has been highlighted by the remodeling of much of the economy to remote work. As a result, as many as 70 percent of companies have suffered losses related to data breaches. As experts predict, the pandemic situation will affect the implementation in private computers of security measures previously installed in corporate equipment.

– Artificial intelligence modules rely on data analysis, on separating safe behavior from unsafe and non-standard behavior. In other words, any behavior that can be considered dangerous and can be detected by a human can also be detected by an artificial intelligence algorithm based on the same data. However, the algorithm will do it much faster – argues Sebastian Stefanowski, director of the Practice Department at Inetum Poland, in an interview with the Newseria Innowacje news agency.

Artificial intelligence is capable of processing billions of data records, coming from both databases and unstructured sources, such as blogs. AI learns to recognize threats through machine learning. It gathers information and then determines the threat profile found in, for example, malicious files or suspicious IP addresses. As a result, it detects threats up to 60 times faster than a human – the analysis takes only a few seconds.

– The basis of a good artificial intelligence algorithm is data. Collecting data is very often associated with a loss of privacy and is perceived as a certain danger. However, data can be collected anonymously. You also need to be aware that the gains from AI-based security can far outweigh the costs associated with the risk of data loss. Anyone would agree to give data about their behavior to a bank, for example, in order to protect their account – points out Sebastian Stefanowski.

AI algorithms form the basis of the SafeGuard Cyber platform, which provides protection against attacks in social media and digital B2B channels, among others. The platform independently detects incoming malicious files and protects sensitive data from unauthorized access. The solution is used, among others, by cosmetics company Johnson & Johnson, pharmaceutical giants such as Novartis and Abbott, as well as the US Bank of Tampa.

Most companies are already using artificial intelligence modules. It really depends on the maturity level of the company. The basic level is simply filters, e.g. anti-spam or detecting suspicious attachments in IT mail. The next level are various types of network traffic analysis modules. Cloud providers lead in such techniques. Almost all of them have ready-made solutions that can detect and alert on abnormal and dangerous behavior and can make them available to their customers.

However, artificial intelligence is used not only for security design, but also in the preparation of cyber-attacks. It can be used, for example, to continuously monitor the network to detect new vulnerabilities that can be exploited by hackers. AI is also used to manipulate data and also to launch social engineering attacks.

– It’s a constant arms race. Both sides – both the securing side and the side that attacks – use artificial intelligence algorithms. The fact that the pandemic has seen an increase in the number of cyber attacks will force companies to invest more money in security systems, which will certainly entail the development of this industry, i.e. the development of cyber security systems based on artificial intelligence, predicts Sebastian Stefanowski.

The “Data Breach Cost Study Report” shows that the average total cost resulting from data breaches is $3.86 million. The industry most affected by the attacks was healthcare. Interestingly, in the previous year’s report, the cost was slightly higher at $3.92 million. However, as the report’s authors point out, this decrease is only theoretical. The report highlights the problem of a widening gap between the cost of breaches for organizations investing in more advanced security processes, such as automation, and those investing in less expensive security.

The coronavirus pandemic has not been without impact. An IBM report found that the implementation of remote working increased the cost of data breaches by $137,000. This phenomenon was observed by 70 percent of respondents. 76 percent reported that remote working increased the time it took to detect and contain a potential threat.

– The consequence of what happened in 2020 will be that the center of gravity of security thinking will shift from securing computer systems on corporate networks to securing computer systems on private networks. In other words, those artificial intelligence algorithms that are now only available to large companies will become popular and make their way as security features to normal private computers in private homes. This will be a simple consequence of the fact that the majority of employees in some industries have switched to working remotely,” points out Practice Director at Inetum Polska.

Sebastian Stefanowski