Artificial intelligence in IT, trends for 2020

One of the trends for this new decade, artificial intelligence within organizations applied to computer security.

In recent years a trend has emerged in the technological world that has adapted to many areas, being the organizations one of them. Artificial intelligence has helped organizations to reduce operating costs, increase productivity, and offer the customer a better experience when buying. But why are there still organizations that refuse or prefer not to integrate artificial intelligence into their organizations? Well, this is because there is a lack of information about what artificial intelligence is and how it can help them, and the lack of resources to integrate this modality is a factor to be taken into account.

If we examine the companies that have and apply these technologies to improve their services, we can see that there has been an improvement in these companies. However, we must also see that artificial intelligence involves maintaining a database for clients and having all the business information, on the one hand, it is useful for the organization of the companies, but on the other, it can be exposed to cybernetic attacks. Cyber-attacks can range from kidnapping, modifying or eliminating important information for the company to extortion or interruptions in the normal business process, which is why it is necessary to have good cyber-security tools in conjunction with the implementation of artificial intelligence.

To get into the subject, we can define artificial intelligence as “The theory and development of computer systems capable of performing tasks that normally require human intelligence, such as visual perception, speech recognition, decision making and translation between languages” according to the Oxford dictionary. In a simpler view, artificial intelligence consists of a tool capable of simulating some of the human cognitive functions, which are executed on a computer plane. This tool, when taken to an organizational level, this technology is capable of automating some functions that were previously executed by a person, thus increasing efficiency and productivity. Some of these characteristics that artificial intelligence has are: image recognition, improvements in the performance of the commercial algorithm, efficient data processing, object detection and classification, protection against cyber attacks, among others.

Some of the approaches on which artificial intelligence is based have to do with the types of learning, such as automatic learning, which consists of a computer aspect in which the machines have the ability to learn without necessarily being programmed for it, this includes the techniques of supervised and unsupervised learning for prediction applications, analysis and data mining. Automatic learning uses algorithms to learn data patterns so that they can be applied in the future and provide the user with personalized attention; this is the most widely used type of artificial intelligence today
Supervised learning consists of the data used by the algorithm that has been previously tagged, and in this way you can identify information that will later be categorized in the same way it has been taught in the first place. Unsupervised learning is when the algorithm does not use any previous categorization to know how to label the new information, but they do it themselves, so it does not require human intervention.

Another type of learning is the deep, this is not a single algorithm, but a network of them using unsupervised learning, so it involves a more complex process in which it organizes in layers to find complicated patterns among the data, this type of artificial intelligence can be found in technology such as GPS, facial recognition systems, voice recognition, among others.

Due to its wide usefulness, computer security can be well used for the defense of information in terms of information security, as we saw before all the functions that artificial intelligence provides us with, such as automatic learning, data mining, detection and classification, etc; they allow us to design automatic systems that can generate solutions to conflicts without the need of human intervention. Here we find a series of methods used by artificial intelligence in the field of information securit

  • Artificial Neural Networks: These simulate the behavior of human neurons, and work from these to facilitate the prediction of attacks in complex and changing environments.
  • Intelligent agents: This is a program that works independently to meet the needs of a user or another program, in this case it works by communicating with other data to solve a problem related to cyber security.
  • Artificial immune systems: This system also works simulating organic functions of the human being, in this case it imitates the immune functions of a person and takes them to the computer field to be used in computer defense. It uses the characteristics of the biological immune system, such as memory to remember past attacks, more effective response learning against repeated attacks, failure tolerance to recover from system errors and diversity to recognize a wide variety of computer threats and attacks.

Other applications of artificial intelligence in cybersecurity are related to the field of research and the application of resources such as the detection of malicious applications in Android systems. This is research in which several applications were classified in order to detect malware more specifically in devices with the Android operating system, using the previously mentioned artificial immune system and neural networks to produce a program that performs the function of an antivirus, blocking malicious applications and detecting threats. There is also the Voice Identification, this is carried out through a method that is based on reactions for the detection of voice signals, which are compared with existing voice patterns and stored in the database, generally two systems are used, one responsible for recognizing voice patterns and another to generate voice.

Among the research currently being conducted in the field of the use of artificial intelligence in cyber security we have the detection of network vulnerabilities, for this research was conducted a simulation of an evaluative model that favors the situation of a network in its security area, which is done in real time and in a quantifiable way. The results of this research provided data to support the adjustments of the defensive measures.
There are some limitations in the use of artificial intelligence in cyber security, such as the fact that some applications in artificial intelligence require manual updates in the system in terms of knowledge and categorization, due to the fact that we are in a changing world, therefore, the information learned by this system in a specific environment, could not be executed in a very different environment than the one in which such information was conceived.

Also, in the learning stage of these systems, an exhaustive training is required in which several cases are presented so that the system can act on them with the information it has been given, in order to avoid false positives at the time of execution. Likewise, the information that is introduced manually for the process of recognition and classification of the data must be of high quality to avoid these same false positives.

Finally, artificial intelligence can be used to generate attacks, because it works by responding to a stimulus in a given environment, so attackers could evaluate the characteristics of this environment and thus generate new methods to damage the system.

Interesting related articles: Attacking and defending, who wins?

 

Scroll Up