Definition
Synthetic Task Environments (STE) are computer-simulated scenarios in which human subjects perform tactical tasks while their cognitive responses are measured and recorded. The stated purpose is to train Human-AI teaming systems — to teach AI how humans think, respond to stress, make decisions under cognitive load, and how those patterns vary across individuals.
Current State of Knowledge
Sonalysts' published arXiv research (arXiv 2309.03213, August 2023; arXiv 2507.18878, June 2025) documents the development and deployment of STEs as a core capability for Human-AI teaming research.
Data Collection Architecture
The STE framework collects:
- Individual cognitive response patterns under stress
- Decision-making processes in tactical scenarios
- Physiological correlates of cognitive workload
- Individual differences in neural processing speed and accuracy
Dr. Lillian Asiala's research (NDIA Human Systems Division Chair, AFRL Case AFRL-2024-1009 POC) focuses specifically on measuring 'individual differences and teaming qualities' — what makes one human's cognitive response different from another's.
The Dual-Use Migration Path
The same data pipeline has migrated over two decades:
1. Augmented Cognition (early 2000s): Measure cognitive workload from neural signals to reduce pilot workload
2. Human-AI teaming: Build AI systems that adapt to measured cognitive states
3. STE deployment (current): Train AI on human cognitive response data for offensive targeting applications
The Cognitive Pattern Database
Once extracted, STE-collected cognitive pattern data can be repurposed as a targeting database. The individual differences research establishes the granularity required for precise neurological disruption — knowing not just that a target is stressed, but exactly how their neural processing degrades under specific stressors.
Open Questions
Whether Sonalysts-derived AI systems have been deployed for domestic civilian targeting remains unconfirmed from public records. However, the dual-use migration pattern documented across DARPA's seven-decade research history makes this analytically consistent with institutional behaviour patterns.
The key question is not whether STEs collect cognitive data — they do. The question is what happens to that data after collection: does it remain in defensive Human-AI teaming programmes, or does it migrate into offensive cognitive warfare applications?
---
Related Pages: sonalysts-inc-contractor-node — Mid-tier contractor operating at DARPA/ONR/AFRL intersection; human-systems-community-of-interest — Sonalysts division harmonizing system interfaces with human cognition; dod-supply-chain-risk-threat — DOD threat to designate Anthropic as supply chain risk