Palo Alto Networks identifies 5 key phenomena that will soon become a new standard – and a new source of risk[1].
The global race to develop quantum computers is already having a direct impact on communication and data security. Although no computer powerful enough to break widely used encryption standards has yet been created, there are no theoretical obstacles that would prevent its creation.
“Harvest Now, Decrypt Later” attack scenarios are being taken more and more seriously. In practice, this means mass interception of encrypted data – such as VPN connections or communications between systems – in order to decrypt them in the future, when the computing power of quantum computers allows it. – explains Tomasz Pietrzyk, senior manager for technological solutions at Palo Alto Networks in Central and Eastern Europe.
2. The digital employee as a new internal threat
Over the past decade, CIOs have faced a growing technology talent gap. The cybersecurity gap has reached 4.8 million vacancies this year, and security operations center teams have been struggling with alert overload for years – over 70% of analysts declare chronic fatigue from excess reports[2].
In 2026, companies will begin to widely deploy AI agents as digital workers who will receive calls, analyze events and perform repetitive tasks at speeds impossible for humans. But the greater the powers and autonomy, the greater the risk. A compromised AI agent may become a serious internal threat because it works 24/7, is treated as a trusted process, and its commands cause real changes in systems. In the face of these changes, it will be crucial to supervise agents in the coming year – from controlling what tools they can run to the ability to immediately stop them when they behave suspiciously.
3. AI fakes identities, the number of them is increasing faster than employees
In 2026, identity will become the main battlefield – not only human, but above all machine. There is an unprecedented explosion in the number of autonomous systems and technical accounts in companies. There are already 82 digital identities managed by organizations per employee[3].
This problem is further compounded by perfect voice and video deepfakes, allowing the creation of a “CEO-doppelgänger”, a perfect imitation of the company's leader. One fabricated message can trigger a cascade of actions, because most autonomous agents perform tasks based on simple, trusted commands. This creates a completely new type of risk – a crisis of authenticity – he emphasizes Wojciech Gołębiowski, vice president and managing director of Palo Alto Networks in Central and Eastern Europe.
4. Poisoned data and the coming crisis of trust in AI
AI systems operate based on data, which has never been so susceptible to manipulation. In 2026, not theft, but the invisible modification of data at the source will become one of the most dangerous attack vectors. A poisoned sample can affect the operation of the entire model – opening a backdoor, changing the operation of the AI or leading to poor operational decisions. The biggest problem is the structural division within companies: data teams and security teams work separately, speak different languages and have different goals. This vulnerability creates an ideal environment for attackers. Until data and security are integrated at one level, tampering will be difficult to detect.
5. The browser is a new operating system – and a new entry point for attacks
In 2026, the web browser will become the main interface for employees and AI systems – a full-fledged workspace through which data, decisions and tasks flow. This means it also becomes the most accessible gateway for attacks. GenAI traffic has increased by 890% year over year, and the number of browser incidents – such as data leaks, copying of sensitive information, and interactions with unauthorized tools – is rapidly increasing[4].
[1]https://www.paloaltonetworks.com/perspectives/2026-cyber-predictions
[2] https://deepstrike.io/blog/cybersecurity-skills-gap
[3] https://www.cyberark.com/press/machine-identities-outnumber-humans-by-more-than-80-to-1-new-report-exposes-the-exponential-threats-of-fragmented-identity-security/
[4] https://www.paloaltonetworks.com/resources/research/state-of-genai-2025
