Training data ruled out
Courts or lawmakers restrict training on unlicensed work, or order models retrained.
Cuts off 8 boxes, 6 of 53 stages · slows 109 · Show on the map →
Hits directly
Signs
The New York Times’ main copyright claims against OpenAI were allowed to proceed in 2025. Anthropic’s $1.5bn settlement showed courts will price pirated training data, though training on lawfully obtained books was found to be fair use.
What would soften it
Licensing markets mature, and licensed and synthetic data fill the gap.
Cut off 8
Slowed 109
- China critical-mineral export controls
- Training data supply
- US ban on foreign-made robots and drones
- Workflow redesign
- White-collar hiring slowdown
- Cognitive abundance
- Lab automation
- On-prem and air-gapped frontier AI
- Pretraining at scale
- AI-designed medicines
- Conversational assistants
- Materials discovery
- Multimodal perception
- Music generation
- Open-weight models
- RL on checkable tasks
- Strategic AI competition
- Test-time compute
- AI answers replace search
- AI companions
- AI in public services
- AI research tools in investing
- AI tutoring at scale
- AI writes most new code
- Embodied agents
- Frontier lab revenue
- Image generation
- Low-price AI for emerging markets
- Military AI and autonomy
- On-device models
- Perturbation data at scale
- Reasoning models
- Robotaxis at scale
- Sycophancy
- Voice cloning and real-time speech
- World models and simulation
- AI cyber defence
- AI-enabled persuasion
- Ambient AI wearables
- Ambient documentation beyond medicine
- Capex payback gap
- Clinical AI in practice
- Firm-owned post-trained models
- Home robots
- Long-horizon agents
- Narrowing of scientific inquiry
- Opaque reasoning
- Self-replicating robot factories
- Video generation with sound
- Virtual cell models
- AGI: top-expert-level AI
- AI-native medicine
- AI-run cyber operations
- Always-on background agents
- Attention and curation under content glut
- Coding agents as products
- Knowledge-rich agents
- Misbehaviour cascades in agent swarms
- Model weight security
- Offshore services displacement
- Synthetic media at scale
- Trace capture
- Agents that trade for investors
- AGI declared
- AI R&D uplift
- AI-generated trading signals
- AI-native application revenue
- Audit and attribution
- Automated coder
- Compute deterrence
- Computer use
- Continual-learning agents
- Deskilling from AI use
- Expertise pipeline
- Firms shrink to a few operators
- Government AGI project
- Professional-services agents
- Public AI scepticism
- Software that maintains itself
- Verification of AI agreements
- Agent-native organisations
- AI as an electoral issue
- AI in research mathematics
- AI-run capital allocation
- Drop-in remote worker
- Experimental validation capacity
- Institutions for agent societies
- Legitimate governance
- Liability and insurance
- Measured productivity gains
- Multi-agent coordination
- Output verification at scale
- Reputation and credentials
- Rollback
- Scalable oversight
- Superhuman AI researcher
- Verified slowdown
- Agent economies
- Agent-run discovery
- Agents designing agents
- AI-enabled power concentration
- Alignment
- Divergent-copy reconciliation
- Explosive economic growth
- Intelligence explosion
- Internal deployment oversight
- Long-horizon value stability
- Superintelligence
- Long-latency civilisation