About Knoware
What is KnoWare?
KnoWare develops the first scientific methodology to detect and forecast emerging capabilities in General Purpose AI (GPAI) systems —before those capabilities fully materialise— helping organisations, AI providers, and EU regulators stay ahead of AI risks and benefits.
The Core Idea
General Purpose AI models (like GPT-4, Gemini, Claude) develop unexpected new abilities as they scale. These “emerging capabilities” appear seemingly suddenly and are poorly understood. Today, the field reacts after a new capability appears. KnoWare wants to flip this:
From reactive observation
to proactive detection
KnoWare brings Knowledge aWareness to General Purpose AI — shifting capability assessment from reactive observation to proactive detection through a novel prefiguration methodology.
The key concept is “prefiguration” — the detectable early signs that a capability is forming: partial behaviours, latent skills, calibration shifts, and enabling conditions. KnoWare builds tools and frameworks to spot these signs early and assess their associated risks.
prefiguration
emerging
proactive
Project Duration
Partner Organisations
Countries Involved
Key Exploitable Results
KnoWare Solutions
Key Exploitable
Results & Tools
Eight Key Exploitable Results delivered through purpose-built tools, frameworks, and open-source resources — enabling a new paradigm for GPAI governance.

Capabilities Roadmap

Prefiguration Methodology

Open-Source Tool Suite

GPAI Benchmark Suite

Socioeconomic Impact Framework

Environmental Impact Assessment

Regulatory Impact Analysis Framework

Regulatory Alignment Roadmap
Four
interconnected
objectives
KnoWare operates across scientific, technological, societal and legal dimensions — each objective linked to measurable KPIs and Key Exploitable Results.
Prefiguration Methodological Framework
Decision-Support Tools & Benchmark Suite
Impact Assessment Methodology
Regulatory Guidance for Systemic Risks
Energy Trade
• Goal Optimisation
KPLER
GPAI-Powered Decision Support for Energy & Commodity Traders
- ≥5 risky behaviours detected
- ≥80% explainability rate
HEALTH
• Agentic System
University of Guelph
Multimodal Agentic GPAI for Pandemic Modelling & Early Warning
- ≥5 unsafe decisions identified
- ≥75% output explainability
Critical Infrastructure
• Fine-Tuned Models
KPLER / MarineTraffik
Multimodal GPAI for Maritime Cybersecurity Event Detection
- >80% detection accuracy
- 80% events explained by chatbot
Food & Chemical Safety
• Agentic System
BIA Analytical
Agentic GPAI for Food Fraud Detection & Chemical Risk Assessment
- ≥5 risky behaviours detected
- ≥80% explainability rate
Four high-impact
real-world validations
KnoWare’s methodology and tools are co-designed and validated through industrial use cases in high-risk areas with significant economic potential and societal implications.
Expected Impact
Results that reframe GPAI governance
KnoWare delivers measurable scientific, technological, societal, and regulatory outcomes that advance Europe’s leadership in trustworthy AI.
GPAI Capabilities Typology & Prefiguration Framework
Open-Source
Suite & Novel Benchmarks
Socioeconomic & Environmental Impact Methodologies