In 2026, will the world be hit by an unprecedented cyber attack because of AI manipulation?

Fortinet's report predicts that cybercriminals will exploit AI vulnerabilities to distort models, create highly realistic deepfakes, and expand attacks to GPS systems, satellites, and quantum encryption.

More sophisticated threats are targeting the integrity and reliability of AI models directly.

Accordingly, by 2026, the field of cybersecurity will become more complex, with the biggest threat coming from the "maturity" of cyber attacks using artificial intelligence (AI).

Attack with new technology from AI

According to Fortinet, the development of generative AI has "democratized" technology, bringing high efficiency but at the same time giving control to more people, increasing security risks.

AI models lack transparency, and uploading sensitive data to the cloud increases the potential for leaks of personal information and intellectual property.

In particular, malicious actors are developing sophisticated methods to exploit vulnerabilities in AI.

Forms of model attacks include manipulating input data to trick the AI ​​into making false predictions, data “poisoning” attacks, and the most dangerous form, Prompt Injection – embedding hidden instructions that cause the AI ​​to bypass safety rules or execute malicious commands.

By 2026, AI models will have access to more sensitive data and automated communication between technology agents, enabling large-scale attacks.

Deepfake and high-level psychological manipulation.

Along with that, Deepfake is predicted to push fraud and psychological manipulation (social engineering) to a new level. AI has the ability to create sounds, images and videos that are almost impossible to distinguish between real and fake.

Deepfake-as-a-Service will be widely used to carry out highly targeted attacks such as business email compromise (BEC), phishing, and extortion, forcing organizations to confront disinformation in a much more difficult-to-control way.

Additionally, the simplification of espionage by AI, combined with the growth of “service” and sponsored cybercrime groups, will continue to fuel attacks targeting multi-billion dollar multinational organizations.

New attack targets

Alongside AI, emerging technologies are also creating new targets for attackers. Space-based GPS systems have been identified as a prime target, where bad actors could jam, block or spoof signals to undermine critical infrastructure.

The explosion of satellite internet also comes with risks, as unencrypted data can be easily intercepted using simple receivers. At the same time, quantum computing, while not an immediate threat, remains a serious warning; quantum computers can break almost any existing encryption, allowing attackers to employ a "collect data first, decrypt later" strategy while waiting for the technology to mature.

In the context of increasingly dangerous cyberattacks, cybersecurity managers need to change their old ways of working. They need to be more than just technical experts; they need to become strategists, transforming security into a factor that helps the company grow and become more trustworthy. They need to leverage AI to quickly detect and stop attacks. More importantly, they must rigorously protect the company's AI systems.

Back to topic

EAST SEA

Source: https://tuoitre.vn/nam-2026-the-gioi-se-bi-sieu-tan-cong-mang-chua-tung-co-vi-ai-bi-thao-tung-20251208160719164.htm