Human-in-the-loop review
Glossary Review Desk
Review the complete term contract without editing the source. Decisions save only in this browser and export as a version-bound JSON checkpoint; they do not publish or change a term automatically.
Needs review
17 terms
Reviewed in POC
217 terms
- Actor
- Adversarial Review
- AI Agent
- AI Candidate
- AI Governance
- AI Literacy
- AI Overreliance
- AI Washing
- Algorithm
- Alignment
- Anchoring
- Android
- Application Programming Interface
- Artificial General Intelligence
- Artificial Intelligence (AI)
- Artificial Superintelligence
- Assumed Isolation
- Audio
- Automated Decision-Making
- Automation Bias
- Automaton
- Autonomous System
- Awareness Boundary
- Batch
- Bespoke / Enterprise Instance
- Bias
- Bottleneck
- Build vs. Buy
- Bullshit, as Frankfurt Defines It
- Cache / KV Cache
- Calibration
- Capability Provenance
- Chain of Thought
- Change Management
- Chaotic Process
- Chatbot
- Chunking
- Client-Safe Language
- Cognitive Offloading
- Communication Gap
- Comprehension Debt
- Compute
- Computer Vision
- Concentration Risk
- Confidence vs. Correctness
- Confirmation Bias
- Constitutional AI
- Constraint Migration
- Content Credentials (C2PA)
- Context Collision
- Context Window
- Context-to-Reality Drift
- Contrastive Learning
- Conversation Drift
- Conversational Inertia
- Copilot
- Copyright / IP Infringement Risk
- Current State
- Curriculum Learning
- Data Leakage
- Data Privacy / GDPR
- Data Provenance
- Data Readiness
- Dataset
- De-Identification / Anonymization Risk
- Debt
- Decision Point
- Deep Learning
- Direct Preference Optimization
- Distributed Commitment
- Distribution Constraint
- Duplicate Entry
- Embodiment Premium
- Epistemic Humility
- Epoch
- EU AI Act
- Evaluation Gap
- Evaluation Metric
- Evidence Selection
- Exception
- Excessive Agency
- Expert System
- Fail-closed
- Fine-Tuning
- Foundation Model
- Friction
- Friction Inventory
- Frontier Model
- Gauntlet Loop
- Generative AI
- Grounding
- Guardrail Displacement
- Guardrail Inversion
- Handoff
- Human Oversight
- Human Ownership
- Human Review
- Human-reserved Work
- Image
- Independent Verification
- Inference-Time Scaling
- Input
- Institutional Memory Cliff
- ISO/IEC 42001
- Jagged Capabilities
- Large Language Model
- Larger Business Context
- Latency
- Latent Representation
- Layered Sovereignty
- Load Rejection
- LoRA
- Loss Function
- Low-Risk Task
- Machine Learning
- Manual Reconciliation
- MCP Server
- ML Target
- Model
- Model Card
- Model Collapse
- Model Context Protocol (MCP)
- Model Risk Management
- Model Routing
- Model Steering
- Module
- Multi-Agent System
- Multimodal
- Multitask Learning
- Natural Language Processing
- Neural Network
- NIST AI Risk Management Framework
- Objective Escape
- Observation
- Open Weights vs. Closed Weights
- Operational Slack
- Orchestration
- Output
- Ownership Gap
- Parameters
- Pattern Triggers
- Persuasion Loop
- Policy Leakage
- Premature Recommendation
- Pretraining
- Process
- Process Context
- Process Debt
- Process Map
- Prompt
- Prompt Engineering
- Prompt Injection
- Prompt or Workflow Chaining
- Proof of Concept
- Race Condition
- Random Seed
- Reasoning Model
- Recommendation Boundary
- Red Teaming
- Relationship State
- Resilience Debt
- Responsible AI
- Retrieval-Augmented Generation
- Rework Loop
- RLHF
- Robot
- Root-Cause Hypothesis
- Runtime Governance
- Safe Curiosity
- Sanitized Information
- Scaling Laws
- Scenario
- Search / Retrieval
- Self-Supervised Learning
- Semantic Debt
- Sensitive Information
- Sensor
- Service-Level Agreement
- Shadow AI
- Small Language Model
- Speech AI
- State
- State Transition Protocol
- Stochastic Process
- Supervised Fine-Tuning
- Supervised Learning
- Sycophancy
- Synchronization Cascade
- Synthetic Data
- System
- System Prompt
- Technical Debt
- Tensor
- Third-Party Audit
- Token
- Tool Use
- Total Cost of Ownership
- Training
- Training Data
- Transformation
- Trust Borrowing
- Unsupervised Learning
- Upstream and Downstream
- Use Case
- Vendor Lock-In
- Vendor-Supported Demand
- Verifiable Domain
- Verification
- Video
- Watermarking
- Weights
- Workaround
- Workaround Dependency
- Workflow
- Workslop
- World Model
- Zero-Shot Learning
Review packet: rs-glossary-glossary-team-poc-0.5.5-2026-09-02 · sha256:684271f1fc45922a57a9af5d42de5436c3f999a5400b3f8065ac53bb6419401d