“Crossing the Rubicon” means passing a point of no return. The idiom comes from Julius Caesar illegally leading his army across the river Rubicon in 49 B.C., an act that sparked the Roman civil war and ultimately made him dictator for life.

But how has cybersecurity crossed the AI Rubicon?

Put simply, the integration of AI into both attack and defense has permanently changed the nature of cybersecurity, creating a before-and-after moment in 2025.

We are witnessing a great acceleration in the speed and scale of change, with an exponential growth in threats, complexity and the deployment of AI tools that characterized the year.

At the same time, cybersecurity has become a geopolitical weapon with a convergence of cyber and real-world conflict. This is a shift from mere data loss to nation-state conflict and hacktivism as the dominant narratives. While ransomware and other cyber attacks continue to grow as cyber themes outside the Washington, D.C., Beltway, new AI-cyber subthemes have emerged. These deeper AI themes from 2025 include:

• **The Agentic Code Tipping Point. **Agentic AI (or code) refers to autonomous AI systems that can reason, plan and execute actions without constant human input. The blogs specifically mention this emerging threat, as it automates and accelerates complex attacks.
• **Adaptive Threat Tipping Point. **This focuses on the effect of AI. It describes attacks that are no longer static but dynamically change tactics in real time to evade defense mechanisms, a key capability that AI provides to attackers.
• **Generative Threat Tipping Point. **This highlights the role of generative AI (like large language models) in creating highly realistic and personalized phishing emails, deepfake scams and massive volumes of attack code, enabling attacks at an unprecedented scale.

In this webinar, my friend and colleague Richard Stiennon explores more perspectives on how AI is dominating cybersecurity advancements and tool sets.

A QUICK LOOK BACK

I always find that a quick recap of the past few years’ summary articles can help put our past 12 months in perspective. Here are my year-end cyber roundups from 2020-2024:

• 2024 Cyber Review: Trump Re-election Eclipses AI and Ransomware Stories — “2024 was a year with growing ransomware attacks, accelerating use of artificial intelligence (for good and evil), more deepfakes in online fraud and election threats being overstated, but Donald Trump becoming president-elect (again) is the top 2024 cyber story.”
• 2023 Cyber Review: The Year GenAI Stole the Show — “This was a year unlike any other in the brief history of the cybersecurity industry, with generative artificial intelligence disrupting plans and ushering in unparalleled change to security.”
• 2022 Cyber Review: The Year the Ukraine War Shocked the World — “This past year will be remembered as another year of ransomware attacks, data breaches impacting critical infrastructure and, most of all, global cybersecurity impacts from the Russian war with Ukraine.”
• 2021 Cyber Review: The Year Ransomware Disrupted Infrastructure — “2021 will be remembered as the most disruptive year so far when it came to cyber attacks, with ransomware impacting businesses and governments — including critical infrastructure — as never before.”
• 2020: The Year the COVID-19 Crisis Brought a Cyber Pandemic — “2020 will be remembered as a uniquely disruptive year — but not just for a global health crisis. Online life was digitally transformed, as exponential change accelerated at home and work via cyberspace.”

OTHER TOP 2025 CYBER THEMES

Here are the top cybersecurity themes from the past year that do not involve AI:

1. Ransomware’s Continued Evolution and Growing Impact Ransomware remains a dominant and brutal threat. Attacks hit a record in 2024 and have not slowed down in 2025 — with another record likely. This theme encompasses:

• Growing Sophistication: Ransomware gangs are evolving, using more sophisticated phishing and social media compromises for initial access.
• Data Exfiltration: A major trend is the increased use of double extortion, where attackers steal data before encrypting it, making the threat of public disclosure a primary leverage point.
• Targeting Critical Sectors: There is a continued and often-cited surge in attacks against critical infrastructure like health care, government, education and water utilities.

2. Cybersecurity of Critical Infrastructure The vulnerability and protection of critical infrastructure is a frequently discussed theme, often independent of AI.

• Specific Attack Examples: Numerous cyber attacks against water facilities, airports and the automotive sector, to name a few.
• Mandated Reporting: The theme includes the move toward mandated reporting requirements for major cyber incidents affecting critical infrastructure operators in the U.S.
• Nation-State Threats: Significant nation-state cyber attacks against critical infrastructure as part of wider global conflicts.

3. Supply Chain Attacks on the Rise The risk posed by third-party vendors and software dependencies is a major non-AI concern.

• Developer Targeting: Attackers are increasingly targeting developers via software package managers and open source ecosystems to compromise the supply chain.
• Impact of Major Incidents: The ripple effects of massive supply chain incidents (like the Change Healthcare or Sisense breaches) underscore the widespread, non-AI-based risks inherent in complex IT ecosystems.

4. Cybersecurity Budgets, Workforces and CISO Roles This theme focuses on the organizational and human elements of cybersecurity.

• Workforce Study: Discussions around the findings of the ISC2 Cybersecurity Workforce Study for 2025, which highlights eye-opening trends in the cyber skills gap.
• Budget Competition: Cybersecurity budgets compete among business priorities in a mixed economic environment and what’s ahead for funding.
• Evolving CISO Role: Chief information security officers have gained more power and a broader role in organizations.

5. Cloud Security Challenges and Evolution While AI can be used in the cloud, the core challenge of securing cloud environments remains a distinct, non-AI theme.

• Hybrid and Multi-Cloud: Attacks targeting hybrid and multi-cloud environments are maturing and becoming more impactful.
• Cloud-Native Threats: The emergence of new attack vectors like cloud-native worm attacks.
• Cloud Security Maturity: The honeymoon for cloud security is now over. There is a shared responsibility model and associated risks with all cloud deployments.

FINAL THOUGHTS

As we head into 2026 and beyond, the total AI dominance within various aspects of cybersecurity will become so obvious and all-encompassing that new terms and subcategories will become necessary to describe the changes and advancements within cybersecurity disciplines and topics. Just like the move from horse and buggy to automobiles for transportation revolutionized the travel industry, AI is radically changing the infrastructure and other aspects of cybersecurity.

For example, if everyone is now driving an automobile, we need new terms to describe the different types of automobiles. In the same way AI cybersecurity tools, which are relevant in both attack and defend scenarios, will make the generic term “AI” almost meaningless.

But as we think back on 2025 in cyber, the AI narrative has reached a point of no return and has fundamentally changed the way we think about cyber attacks, cyber defense and the future of technology protections.

Next week, we explore part 1 of the top 26 security predictions for 2026, which highlights the best cyber industry reports, forecasts, trends and security predictions for 2026 and beyond.

Daniel J. Lohrmann is an internationally recognized cybersecurity leader, technologist, keynote speaker and author.