4 Reasons Why AI Is Dangerous in Phishing Attacks

NGUYỄN ĐĂNG |

The rapid development of AI (artificial intelligence) makes phishing attacks personalized, making it possible for even cybersecurity experts to fall for them.

A recent study by cybersecurity company Kaspersky shows that the number of cyber attacks faced by organizations has increased by nearly 49% in the past 12 months.

The most common threat comes from phishing scams. Half of the survey respondents (50%) expect the number of phishing attacks to increase significantly, as cybercriminals increasingly take advantage of AI.

Cybersecurity experts also pointed out four reasons why phishing scams have become more dangerous and unpredictable due to AI.

Phishing attacks are personalized by AI

Traditionally, phishing attacks relied on blasting out the same generic messages to thousands of people, luring a few into a trap. AI has changed that by creating sophisticated, personalized phishing emails at scale.

AI-integrated tools can use personal public information on social networks, recruitment sites or company websites to create tailored emails that match each individual's role, interests and communication style.

For example, a CFO might receive a phishing email that copies the tone and style of a message from the CEO, even referencing recent company events. This level of customization makes it difficult for employees to distinguish between a real message and a phishing message.

Su phat trien vuot bac cua AI khong chi tac dong den nhieu nganh cong nghiep ma con thay doi chien thuat lua dao cua cac nhom toi pham mang. Anh: Nhien Do
The rapid development of AI not only affects many industries but also changes the fraud tactics of cybercriminal groups. Photo: Nhien Do

AI Makes Deepfakes More Dangerous

Deepfake technology in AI has also become a powerful weapon, used by cybercriminals in scams.

Attackers exploit this technology to create fake audio and video clips, simulating the voices and appearances of leaders and managers with an astonishing level of accuracy.

For example, in one documented case, an attacker used deepfake to impersonate multiple employees in an online meeting, convincing one employee to transfer approximately $25.6 million.

As deepfake technology continues to evolve, attacks of this type are expected to become more widespread and sophisticated.

AI Helps Attackers Bypass Traditional Security Methods

Cybercriminals can use AI to fool traditional email filtering systems. By analyzing and mimicking legitimate email patterns, AI-generated phishing emails can bypass security software checks.

In addition, machine learning algorithms can test and refine scams in real time, increasing the success rate and making scams increasingly sophisticated.

Experts can also fall into AI traps

Even experienced cybersecurity professionals fall victim to sophisticated phishing attacks.

The authenticity and personalization of AI-generated content sometimes overwhelms the skepticism that keeps even experienced professionals on their toes.

Furthermore, these attacks often play on human psychological factors, such as urgency, fear, or power, pressuring employees to act without thoroughly checking the authenticity of the request.

NGUYỄN ĐĂNG
TIN LIÊN QUAN

New path for modern investors

|

2025 is predicted to be a boom year for investment sectors related to technology and innovation in Vietnam. Many “emerging” sectors promise to bring attractive investment opportunities to astute investors.

Google partners with AP to update news in Gemini app

|

Google and the Associated Press (AP) news agency partner to bring real-time news to the Gemini app.

Updated driving licenses on VNeID are slow, Traffic Police Department explains the reason

|

The Traffic Police Department has coordinated with technical units to fix and optimize transmission lines, ensuring quick updating of driving licenses on VNeID.

Hanoi has the first case of Japanese encephalitis of the year, the number of COVID-19 cases doubled in the same period

|

On July 14, the Hanoi Center for Disease Control (CDC) recorded the first case of Japanese encephalitis in 2025, after many months of no new cases.

Billiards legend suddenly passes away while resting at hotel

|

Famous Taiwanese billiards player Chang Jung-lin suddenly passed away while resting at a hotel.

Civil servants wear life jackets, cross the river to the capital's only island commune

|

Hanoi - At 6am, when the sun had not yet risen, the first ferries at Chu Minh Wharf took officers and civil servants across the Red River to Minh Chau island commune.

2 apartments in the middle of the city owe more than 400 red books to people in Quang Ninh

|

Quang Ninh - More than 400 households have moved to both apartment buildings in Hong Gai ward, Quang Ninh province for a long time but have not yet received their red books.

Look up the fastest high school graduation exam scores in 2025 on Lao Dong

|

The 2025 high school graduation exam scores will be announced by the Ministry of Education and Training at 8:00 a.m. on July 16. Candidates can look up the fastest exam scores on Lao Dong.

New path for modern investors

Lục Giang |

2025 is predicted to be a boom year for investment sectors related to technology and innovation in Vietnam. Many “emerging” sectors promise to bring attractive investment opportunities to astute investors.

Google partners with AP to update news in Gemini app

TRÍ MINH (THEO phonearena) |

Google and the Associated Press (AP) news agency partner to bring real-time news to the Gemini app.

Cảnh báo về các hình thức tấn công giả mạo để vượt xác thực 2 yếu tố

NGUYỄN ĐĂNG |

Các chuyên gia an ninh mạng của Kaspersky đã phát hiện ra các hình thức tấn công giả mạo (phishing) được tội phạm mạng sử dụng để vượt xác thực 2 yếu tố (2FA) – vốn là một biện pháp bảo mật quan trọng được thiết kế để bảo vệ tài khoản trực tuyến.