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 enabling less skilled actors to conduct sophisticated cyberattacks, disrupting established threat evaluation. The old metrics no longer reliably identify dangerous attackers.
Anthropic’s latest analysis confirms that AI is significantly increasing the danger level of cyberattackers while rendering traditional threat assessment metrics ineffective.
Between March 2025 and March 2026, Anthropic examined 832 banned accounts involved in malicious cyber activity, finding that AI is increasingly used to prepare and execute attacks. The majority of these actors employed AI to generate malware, with 67.3% of cases involving AI-assisted malware development. Moreover, the use of AI for lateral movement within compromised networks rose from 33% to 56% over the year, indicating attackers are deploying AI deeper into attack phases.
Importantly, the report shows that AI enables less skilled actors to perform complex, previously skill-dependent tasks such as account discovery and lateral movement. This democratization of capabilities means threat actors with minimal technical expertise can now carry out more dangerous operations, challenging existing threat models that rely on the number of techniques or tools used to assess risk.
The frameworks can’t see the thing that matters — ThorstenMeyerAI.com
<!– DEPLOY: swap this Google Fonts for self-hosted base64 woff2 (German GDPR) –>
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
“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.
⌁ there is no MITRE ATT&CK ID for agentic orchestration
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.
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.
dead signal
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.
fading signal
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.
durable signal
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.
Impact of AI on Threat Detection and Risk Assessment
This development fundamentally alters how cybersecurity professionals evaluate threats. Traditional heuristics—such as counting techniques or analyzing tools—are no longer reliable indicators of an actor’s danger level. As AI enables less skilled actors to perform advanced techniques, the threat landscape becomes more unpredictable and harder to assess, increasing the risk of underestimating threats.
The report emphasizes that the ability to identify high-risk actors based on their use of complex techniques or specific tools is diminishing. Instead, the focus shifts to how and when attackers deploy AI during different attack stages, which is a more subtle but critical indicator of threat level.
Evolution of Cyberattack Techniques and Threat Models
Historically, threat assessment relied on the assumption that more techniques and sophisticated tools indicated higher danger. This heuristic was supported by decades of security practice, which linked attacker skill directly to the complexity of techniques used. However, recent advances in AI, particularly large language models and automation tools, have changed this landscape.
Anthropic’s analysis builds on previous reports, including Verizon’s 2026 Data Breach Investigations Report, highlighting a shift towards AI-driven attack preparation and execution. The trend reflects broader technological developments, with AI lowering barriers for less experienced actors to carry out high-impact attacks.
“Our data shows a clear shift towards AI being used not just for initial access, but for deeper, more operational phases of cyberattacks, which was previously restricted to highly skilled actors.”
— Anthropic report author
Unclear Aspects of AI’s Long-Term Impact on Threat Dynamics
It remains uncertain how quickly threat assessment frameworks will adapt to these changes or whether new models will emerge that can reliably differentiate between high- and low-risk actors in an AI-enabled landscape. The full scope of AI’s influence on future attack sophistication and detection remains to be seen, as ongoing developments could accelerate or alter current trends.
Future Steps for Cybersecurity Threat Evaluation
Cybersecurity agencies and organizations are expected to reevaluate threat assessment protocols, incorporating new signals such as attack timing and operational context. Further research and real-time monitoring will be necessary to understand how threat actors evolve with AI and to develop adaptive defense strategies. The industry may also see increased emphasis on AI-specific detection tools and threat intelligence that account for AI-driven attack techniques.
Key Questions
How does AI change the way attackers operate?
AI enables attackers to automate complex tasks like lateral movement and account discovery, making less skilled actors capable of executing sophisticated attacks.
Why are traditional threat assessment methods no longer effective?
Because AI allows even low-skill attackers to perform techniques that previously required expertise, the correlation between skill level and technique count or tool use has broken down.
What are the implications for cybersecurity defenders?
Defenders need to develop new detection strategies that focus on attack behavior and operational signals rather than just techniques or tools used.
Is this trend likely to accelerate?
Yes, as AI technology becomes more accessible and advanced, the trend of democratized attack capabilities is expected to continue, further complicating threat assessment.
What can organizations do to prepare?
Organizations should invest in AI-aware detection tools, update threat models, and enhance threat intelligence to better identify AI-enabled attack behaviors.
Source: ThorstenMeyerAI.com