📊 Full opportunity report: The Frameworks Can’t See the Thing That Matters: A Year of AI-Enabled Cyber Threats on ThorstenMeyerAI.com — validation score, market gap, and execution plan.
TL;DR
A year-long analysis shows AI is empowering less skilled attackers to perform complex malicious activities, challenging existing threat assessment methods. The traditional focus on techniques and tools no longer reliably indicates threat level.
A new analysis from Anthropic shows that AI is significantly increasing the danger posed by cyber attackers, with traditional methods of threat assessment becoming ineffective. The report highlights how AI-driven activities are enabling less skilled actors to carry out complex attacks, fundamentally changing the landscape of cyber threats and risk evaluation in 2026.
Anthropic examined 832 accounts banned for malicious activity between March 2025 and March 2026, mapping their techniques onto the MITRE ATT&CK framework. The findings reveal that AI is primarily used to automate the creation of attack tools, such as malware, with 67.3% of actors employing AI for this purpose. Notably, a smaller but critical segment used AI for advanced tasks like lateral movement within networks, with the proportion of high-risk actors rising from 33% in the first half of the year to 56% in the second.
Furthermore, the report indicates a shift in AI usage from initial access techniques, like phishing, toward post-compromise activities such as account discovery and lateral movement. This shift suggests attackers are leveraging AI to deepen their infiltration once inside a network, making attacks more dangerous and accessible to less sophisticated actors. The traditional markers of threat—technique diversity and tool sophistication—no longer reliably distinguish high-risk actors, as even less skilled attackers now demonstrate comparable technical breadth due to AI assistance.
The frameworks can’t see the thing that matters
For decades, danger meant which techniques an attacker commands. A year of real AI-enabled attacks — 832 banned accounts mapped onto MITRE ATT&CK — shows that signal breaking, just as a new, harder-to-see one takes over.
A year of real misuse, mapped to the standard taxonomy
A window, not a census — these are the cases with enough detail to assess techniques thoroughly. Inside it, the risk level climbed fast.
WHAT WAS STUDIED
THE RISK CLIMB · MEDIUM-OR-HIGHER ACTORS
Python Scripting for Cybersecurity: Linux Edition: Volume 2 – Log Analysis, Network Visibility, and Threat Detection with Hands-On Python Projects
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
“More techniques” stopped meaning “more dangerous”
The old heuristic: count the techniques, judge the tooling. AI dissolved it — because the model supplies the techniques either way. Watch the old signal fail, then watch what it misses.
Risk score vs. technique count
Two ways to read the same attacker. One is going blind. Press play.
OSINT 2.0: AI-Powered Open-Source Intelligence for Beginners (OSINT 2.0 — Artificial Intelligence for Open-Source Intelligence and Cyber Investigations Book 1)
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Deeper into the attack — and into less-skilled hands
Across the year, AI use drifted from getting in toward acting once already inside — the operationally demanding stages that used to require an expert.
The attack lifecycle · where AI is now applied
The center of gravity moved right — toward post-compromise work.
Network Intrusion Detection
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
From “what they know” to “what they’ve built”
The report sorts the signals into three tiers — one dead, one fading, one durable.
Technique count & tooling
16 vs. 20 between novice and expert; platform doesn’t correlate. The model supplies the techniques either way.
Where in the lifecycle AI is applied
Concentrating on operationally demanding, post-compromise stages is a better signal — but it’s eroding as the whole population heads there.
The scaffolding around the model
Architectures that let the model chain stages and run with minimal human input. Not what they know — whether they’ve built a system that lets AI run the attack.
Operationalizing Threat Intelligence: A guide to developing and operationalizing cyber threat intelligence programs
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Fixing the map before the territory moves again
A taxonomy that can’t name the most dangerous behavior on the field will quietly mislead the people relying on it. The response runs in two directions.
Fed back into the models
The findings informed safeguards on the most capable models, built to detect & block some of what was observed:
- Blocking malware development
- Blocking mass data exfiltration
- Putting tools in defenders’ hands first (Project Glasswing)
Taking it to the source
Following the Verizon work, Anthropic says it’s in discussions with MITRE about how ATT&CK might evolve:
- A vocabulary for agentic orchestration
- Naming the scaffolding that makes a model an operator
- An interactive technique visualization on the Red blog
Reading it in proportion
- The 832 cases are a detailed subset, not the full population — the precise percentages are directional, not definitive.
- “More autonomous” is not “fully autonomous” — even the standout case needed human input at key moments, which is itself a place for defenders to intervene.
- This is one vendor’s window — the company with visibility into misuse of its own model, publishing what it found. The right thing to do with the data, and worth remembering as you read it.
Implications of AI-Driven Attackers on Threat Assessment
This development fundamentally alters how cybersecurity professionals evaluate threats. The reliance on the number of techniques or the sophistication of tools as indicators of attacker danger is no longer valid. AI democratizes advanced attack capabilities, enabling less skilled actors to perform complex operations previously limited to experts. Consequently, organizations may underestimate threat levels if they continue to use traditional heuristics, increasing the risk of breaches.
Evolution of Cyberattack Techniques and AI’s Role
Historically, threat assessment depended on the assumption that more techniques and better tools equated to higher danger. The MITRE ATT&CK framework served as a standard for categorizing attacker tactics. However, recent developments show that AI now automates and simplifies complex attack steps, reducing the correlation between attacker skill and observable activity. This shift has emerged over the past year, driven by advances in large language models and automation tools that lower technical barriers for malicious actors.
“AI is transforming the threat landscape by enabling less skilled actors to perform sophisticated attacks, rendering traditional threat indicators obsolete.”
— Thorsten Meyer, AI security researcher
Uncertainties About the Full Scope of AI-Enabled Threats
It remains unclear how widespread AI-enabled attack techniques are beyond the subset analyzed, and whether threat actors will adopt new AI tools at a faster rate. Additionally, the long-term impact of AI on attack sophistication and threat assessment frameworks is still being evaluated. The current data provides a snapshot but not a comprehensive picture of the evolving threat landscape.
Monitoring and Updating Threat Assessment Strategies
Cybersecurity teams are expected to adapt by developing new detection methods that focus on behavioral signals and attack scaffolding rather than technique diversity. Continued research and real-time monitoring of threat actor behavior will be essential to keep pace with AI-driven changes. Further analysis of attack patterns and AI’s role in operational stages will inform future threat models and defense strategies.
Key Questions
How does AI make attackers more dangerous?
AI automates complex attack tasks, allowing less skilled actors to perform sophisticated operations like lateral movement and account discovery, which previously required expertise.
Why can’t traditional threat assessment methods identify high-risk attackers?
Because AI enables attackers to perform a wide range of techniques regardless of their skill level, making technique count and tool sophistication unreliable indicators of threat level.
What should organizations do to improve threat detection?
Organizations should focus on behavioral analysis, attack scaffolding, and operational signals rather than solely relying on technique diversity or tool type.
Is this trend already widespread or still emerging?
The analysis is based on a subset of attacks over the past year, indicating a significant trend, but the full extent of AI’s role in cyber threats is still unfolding.
What are the risks of underestimating AI-enabled threats?
Underestimating these threats could lead to insufficient defenses, making organizations more vulnerable to breaches carried out by less skilled but highly capable attackers.
Source: ThorstenMeyerAI.com
