Comparative Analysis of AI Software Against Growing Malware Threats
Lehtonen, Ville (2025)
2025
All rights reserved. This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2025052817459
https://urn.fi/URN:NBN:fi:amk-2025052817459
Tiivistelmä
In the information technology sector, malicious software (malware) has the potential to cause considerable damage to the system. Traditional malware detection tools may be inadequate for this challenge. The purpose of the thesis was to find out what artificial intelligence (AI) can provide in terms
of malware defense and detection, by addressing three main research questions. The first research question focuses on the non-AI methods for detecting malware inside of a system. The second question explores the potential AI has in malware detection as well as its shortcomings. Lastly, the third question was about the ethical aspect of these methods, and why the user data is being collected. The preceding results of the thesis are intended to give concrete information to individuals working with cybersecurity.
Both beginners and more experienced individuals could find these results useful.
The topic of this thesis was approached from a theoretical standpoint, focusing on analyzing and reviewing past literature and publications that are related to malware detection, AI, and cybersecurity. First, the research covered cybersecurity in recent years, its motivations and real-life examples. Second, the fundamental concepts of malware were explained. Thirdly, the thesis went into the ethical and sustainable aspects of malware detection tools. Finally, the methodology of this thesis was discussed. The primary method for the thesis was a comparative analysis of available software. The malware detection tools were evaluated in terms of efficiency, potential, data practices, and accessibility.
The research and analysis demonstrated the potential that AI-based malware detection tools could harness. The usage of AI enables the software to detect more sophisticated threats, which more traditional methods have been known to struggle with. This, however, comes with a cost, as using AI directly inside of the software requires a larger amount of system resources than usual. In addition to this, most of the software lacked greater transparency in terms of data privacy. This leaves room for improvement in terms of clarifying the data usage to the user, as well as making the software itself more
lightweight for the systems.
The study indicates that, while AI dramatically improves malware detection, its implementation must meet resource demands and ethical concerns. This is especially important regarding data privacy, which should benefit both newcomers and experienced cybersecurity practitioners. These findings highlight the significance of building AI-powered systems that are lightweight in resource usage and transparent with data usage.
Future works could combine AI’s flexible and powerful nature with the reliability and transparency of traditional tools. This approach could improve overall cybersecurity across a multitude of different devices and systems. Additionally, future solutions should implement better and clearer data collection policies. Future research should evaluate the user’s willingness to share data for security purposes, along with testing malware detection tools in different digital environments, such as smart homes.
of malware defense and detection, by addressing three main research questions. The first research question focuses on the non-AI methods for detecting malware inside of a system. The second question explores the potential AI has in malware detection as well as its shortcomings. Lastly, the third question was about the ethical aspect of these methods, and why the user data is being collected. The preceding results of the thesis are intended to give concrete information to individuals working with cybersecurity.
Both beginners and more experienced individuals could find these results useful.
The topic of this thesis was approached from a theoretical standpoint, focusing on analyzing and reviewing past literature and publications that are related to malware detection, AI, and cybersecurity. First, the research covered cybersecurity in recent years, its motivations and real-life examples. Second, the fundamental concepts of malware were explained. Thirdly, the thesis went into the ethical and sustainable aspects of malware detection tools. Finally, the methodology of this thesis was discussed. The primary method for the thesis was a comparative analysis of available software. The malware detection tools were evaluated in terms of efficiency, potential, data practices, and accessibility.
The research and analysis demonstrated the potential that AI-based malware detection tools could harness. The usage of AI enables the software to detect more sophisticated threats, which more traditional methods have been known to struggle with. This, however, comes with a cost, as using AI directly inside of the software requires a larger amount of system resources than usual. In addition to this, most of the software lacked greater transparency in terms of data privacy. This leaves room for improvement in terms of clarifying the data usage to the user, as well as making the software itself more
lightweight for the systems.
The study indicates that, while AI dramatically improves malware detection, its implementation must meet resource demands and ethical concerns. This is especially important regarding data privacy, which should benefit both newcomers and experienced cybersecurity practitioners. These findings highlight the significance of building AI-powered systems that are lightweight in resource usage and transparent with data usage.
Future works could combine AI’s flexible and powerful nature with the reliability and transparency of traditional tools. This approach could improve overall cybersecurity across a multitude of different devices and systems. Additionally, future solutions should implement better and clearer data collection policies. Future research should evaluate the user’s willingness to share data for security purposes, along with testing malware detection tools in different digital environments, such as smart homes.