AI in Cybersecurity: A Double-Edged Sword

As society becomes increasingly digitized, cybercrime is surging and AI is emerging as a key force reshaping the cybersecurity landscape. On the defensive side, AI excels at automating analysis, detecting anomalies at scale, and enabling faster, more proactive responses that strengthen the resilience of digital systems. However, the same capabilities also empower attackers through more convincing social engineering, deepfakes, and AI‑assisted phishing or malware, so organizations and individuals must pair AI adoption with strong awareness, training, and sound security practices to reduce human‑factor risk.

Written by

FPT Software

Published on

August 23, 2024

As we move deeper into the Information Age, digital technologies have become an integral part of modern society. However, as technology continues to advance and our reliance on it grows, constantly evolving cybersecurity threats have become a major concern.

With the rapid introduction and adoption of Artificial Intelligence (AI), an important question arises: to what extent will AI reshape and influence the overall cybersecurity landscape?

The growing threat of cybercrime

As digital transformation reshapes the global economy, everyday activities are increasingly dependent on digital platforms. This deep reliance on technology has dramatically expanded the attack surface, making exposure to cybersecurity risks a critical concern for individuals, businesses, and governments alike.

According to the Identity Theft Resource Center's annual data breach report, there were 3,205 recorded data compromises in the United States in 2023, affecting more than 300,000 victims [1]. Cybersecurity breaches can cause severe damage, including financial losses, leakage of sensitive information, data tampering or destruction, and significant service disruptions. A recent example is the compromise of the Indian Council of Medical Research, which exposed the personal data of 81.5 million Indian citizens and became one of the largest cybersecurity incidents of 2023 [2].

The rate and impact of cybercrime are expected to accelerate as technology becomes even more intertwined with both professional and personal life. From 2018 to 2023, the estimated annual cost of cybercrime rose more than ninefold, reaching 8.15 trillion U.S. dollars. Projections suggest this figure could climb to 13.82 trillion U.S. dollars by 2028 [3]. It is also important to recognize that these statistics reflect only reported incidents and likely underestimate the true scale of cybercrime.

In parallel with rapid technological advancement, ensuring adequate cybersecurity to protect the availability, integrity, and confidentiality of systems and data has become an urgent challenge that must be addressed.

The emergence of AI in cybersecurity

Cybersecurity investigations generate massive volumes of information at every stage, both before and after incidents. As the number of cases continues to rise, organizations inevitably face workforce shortages and human error. In this context, the power of AI becomes crucial: with strong capabilities in automation and data analysis, AI can substantially increase efficiency in handling large-scale information.

One of the most prominent applications of AI in cybersecurity is threat detection. Effective prevention and response require continuous monitoring to identify anomalies. AI models can be trained to understand the normal state of a system and to flag events that deviate from this baseline. Once trained, such models can raise anomaly alerts within microseconds. This approach is used to identify malicious behavior patterns in networked systems, for example through unusual traffic flows or abnormal log alerts. In environments where many machines operate and communicate simultaneously, AI outperforms human analysts in processing high-volume data streams. While a human expert may be able to examine up to 12,000 events per day, AI can accelerate this work by a factor of 1,000, reviewing as many as 12 million events in the same time frame [4].

This capability is now widely adopted to address threats such as malware, fraud, network attacks, phishing, and more, thereby improving the overall security of systems and networks. Applying the same logic, AI models learn the normal behavior of a program, an operation, an email, or a transaction, and detect when these entities behave differently or contain unusual content. Integrating AI into cybersecurity enables a more proactive approach to threat intelligence by uncovering hidden problematic patterns at an early stage. AI-powered solutions can also go further by pinpointing abnormal areas within the system, supporting human investigators in responding to incidents more quickly. Companies such as Microsoft Azure have leveraged AI solutions to reduce threat analysis and detection times from months to minutes, while simultaneously lowering both downtime and recovery costs [5].

A double-edged sword

The growing positive influence of AI in cybersecurity is undeniable, but it is not an inherently trustworthy ally. As with any technology, its impact depends on the intentions of the user. Because AI has become widely accessible to the general public, cybercriminals and hackers can now harness the same capabilities to support their attacks.

One of the clearest examples of AI being misused is in social engineering, where attackers exploit human error to compromise users and systems. With the rise of generative AI, fake text, video, and audio can be produced at the press of a button, enabling highly convincing deceptive content that can fool even experienced users. Hackers can use AI to carry out realistic impersonation, identity theft, and manipulation [6]. Deepfake scams, where close friends or family members are convincingly imitated, are becoming increasingly common, with reported cases rising by 3000% in 2023 alone [7]. As AI models become more sophisticated, this threat is expected to escalate further in the coming years.

Since the launch of GenAI platforms, many actors have attempted to exploit the model's capabilities for malicious purposes. Although the systems include ethical and safety checks, these GenAI models can still be misused to draft phishing emails or SMS messages and to generate harmful scripts for attackers. Research indicates that AI-generated phishing emails are only slightly less effective (21%) at attracting clicks from victims than human-written emails (27%), yet they can be produced at least 40% faster. This efficiency significantly increases the overall success rate of phishing campaigns [8].

Security alongside AI

As AI technology rapidly advances, the cybersecurity landscape is evolving in tandem. AI is set to remain a key actor in this field, helping digital systems strengthen their defenses through more efficient data processing, faster response times, continuous monitoring, and reduced costs and manual workload.

At the same time, both organizations and everyday users must stay alert to cybersecurity threats enabled by AI, which often target the human element. Raising awareness, investing in cybersecurity training, and maintaining sound business practices are all critical to preserving strong security while minimizing the risk of human error.

Frequently Asked Questions

How is digital transformation changing cyber risk today?

As businesses and daily life move onto digital platforms, the attack surface expands and breaches become more frequent and costly. AI will increasingly shape this landscape by automating both defense and attacks, changing how threats emerge, are detected, and are responded to across industries.

Why is cybercrime growing so fast and becoming so costly?

Cybercrime is surging because more valuable data and services are online, attackers are better organized, and many incidents go undetected or unreported. The result is massive financial loss, data exposure, system disruption, and long-term reputational damage that is forecast to keep rising sharply.

How is AI being used to strengthen cybersecurity defenses?

AI helps security teams handle huge data volumes by automating monitoring, anomaly detection, and investigation support. Trained models learn what “normal” looks like, flag deviations in microseconds, and highlight likely attack points—cutting analysis time from months to minutes and reducing costs and human error.

In what ways can attackers misuse AI to improve their attacks?

Attackers use accessible AI tools to craft convincing phishing emails, deepfake voices and videos, and tailored malicious scripts faster and at scale. Generative models enable realistic impersonation and social engineering, boosting click rates and lowering the skills needed to launch impactful cyberattacks.

How can organizations run secure operations in an AI-driven world?

Organizations should deploy AI for continuous monitoring, rapid detection, and faster response while investing in user awareness, security training, and strong governance. Combining AI-driven analytics with clear processes and educated staff helps cut costs and human error while adapting to evolving AI-enabled threats.