📊 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 increasingly used by cybercriminals to enhance attack complexity and scale. Traditional threat indicators no longer reliably distinguish high-risk actors. This shift raises concerns about current cybersecurity defenses.
A new analysis from Anthropic indicates that AI is significantly altering the landscape of cyber threats, making attackers more capable and harder to identify using traditional metrics. The report, based on 832 banned accounts, shows that AI is increasingly used for both mundane and complex attack preparations, challenging established threat assessment models.
Anthropic examined 832 accounts banned for malicious cyber activity between March 2025 and March 2026, mapping their techniques onto the MITRE ATT&CK framework. The analysis reveals that 67.3% of these actors used AI to prepare for attacks, primarily for malware development. A notable trend is the shift of AI use from initial access techniques to post-compromise activities like lateral movement and account discovery, with these activities rising significantly over the year.
Importantly, the report finds that traditional indicators—such as the number of techniques used or the tool interfaces—no longer reliably distinguish high-risk actors. Both novice and skilled actors now employ similar numbers of techniques, often supplied or supported by AI models, blurring the lines of threat classification. Instead, the most dangerous actors focus AI on operationally demanding tasks, but this signal is also weakening as more actors adopt similar approaches.
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
Artificial Intelligence for Cybersecurity: How AI Detects Cyber Threats, Prevents Hacking, and Protects Your Data, Identity, and Smart Devices (AI Cybersecurity Mastery Series)
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.
Industrial Network Security: Securing Critical Infrastructure Networks for Smart Grid, SCADA, and Other Industrial Control Systems
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.
Practical Malware Analysis: The Hands-On Guide to Dissecting Malicious Software
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.
cyber attack simulation kits
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 Attack Evolution for Cybersecurity
This development fundamentally challenges existing threat assessment methods, which rely on technique diversity and tool sophistication to gauge attacker danger. As AI democratizes complex attack capabilities, defenders face the risk of underestimating threats or misallocating resources. The shift toward deeper, post-compromise activities suggests that attackers can now operate more stealthily and effectively, even if they lack extensive technical expertise. This change underscores the urgent need for new detection strategies that go beyond traditional heuristics.
Rise of AI in Cyberattack Tactics and Historical Threat Models
For decades, cybersecurity professionals assessed threats based on the variety of techniques and tools used by attackers. The MITRE ATT&CK framework has served as a standard for categorizing tactics, enabling defenders to identify and prioritize threats. However, recent developments show that AI is enabling less skilled actors to perform complex tasks previously reserved for highly skilled hackers. This evolution aligns with broader trends of AI democratization but poses new challenges for threat detection and response.
“Our findings indicate that AI is not just a force multiplier but is fundamentally changing who can pose a threat and how they operate. The old metrics no longer reflect the real danger.”
— Thorsten Meyer, lead researcher at Anthropic
Unclear Extent and Future Trajectory of AI-Enabled Threats
While the report provides strong evidence of AI’s role in current attacks, it is unclear how widespread these practices will become over the next year. The data is limited to a subset of cases with sufficient detail, and the full scope of AI’s impact remains to be seen. Additionally, the pace of technological development and potential countermeasures are still evolving, making future threat levels uncertain.
Monitoring and Developing New Defense Strategies
Cybersecurity organizations are expected to invest in advanced detection tools that can identify AI-supported behaviors and post-compromise activities. Researchers and practitioners will likely focus on developing frameworks that do not rely solely on technique counts or tool interfaces. Ongoing analysis of attack patterns and AI’s role in threat evolution will be critical in adapting defenses and policy responses.
Key Questions
How does AI change the way cyber threats are assessed?
AI allows attackers to perform complex tasks, such as lateral movement and account discovery, with less skill and effort. This makes traditional metrics like technique diversity less effective for threat assessment.
Are traditional cybersecurity tools still effective against AI-enabled attacks?
Many existing tools focus on detecting known techniques and tool signatures, which are less reliable as attackers use AI to perform operations more stealthily and adaptively. New detection strategies are needed.
What can organizations do to prepare for AI-empowered cyber threats?
Organizations should invest in AI-aware detection systems, enhance monitoring of post-compromise activities, and develop threat models that account for AI-supported attack techniques.
Will AI make all cyberattacks more dangerous?
Not necessarily. While AI lowers the skill barrier for attackers and enables more complex operations, the overall threat landscape depends on how quickly defenders adapt and develop countermeasures.
Source: ThorstenMeyerAI.com
