About LEWS

Team information and background on the Lock-in Early Warning System

Project Background

LEWS (Lock-in Early Warning System) was developed as part of the EA Animal Welfare Hackathon 2025. The project aims to create a simple tool that detects when emerging animal-related technologies (like insect farming) are approaching irreversible lock-in — using early-warning signals inspired by historical patterns.

Project Origin

The project originated from the recognition that emerging animal farming technologies like insect farming and AI shrimp aquaculture need early monitoring to prevent permanent suffering for billions or trillions of animals. The team recognized a gap: no tool existed to predict technological lock-in for animal systems.

Team Roles & Contributions

Project Lead

Responsibilities:

  • Provided the conceptual framework
  • Handled research background
  • Prepared all structured data
  • Created datasets for insect farming, AI shrimp systems, wildlife AI, and factory farming
  • Final UI/UX decisions
  • Presenting the project

Abid (Full Stack Developer)

Responsibilities:

  • UI development
  • Score logic implementation
  • Slider interactions
  • JSON integration

Tim (Frontend Developer)

Responsibilities:

  • Layout polishing
  • Animations and transitions
  • Icon implementation
  • Consistency across components
  • Provided component logic

Thomas (AI/ML Specialist)

Responsibilities:

  • Validation of scoring models
  • Uncertainty handling
  • Data normalization
  • Retrodiction model (optional)
  • Pattern matching algorithms

Rapha (Sociology/Ethics Specialist)

Responsibilities:

  • Conceptual oversight
  • Ethical framework development
  • Critique of model assumptions
  • Advise on advocacy and governance
  • Narrative calibration for judges

Development Approach

The team followed an agile development approach typical for hackathons:

  • Preparation: All data, definitions, and example trajectories were prepared in advance
  • Simple Tech Stack: Node.js backend with Express, JSON data storage, and React frontend
  • Clear Division: Research and technical roles were clearly separated
  • Integration Goal: Complete integration by Hour 15 of the hackathon

Conceptual Foundations

LEWS is grounded in several academic and effective altruism frameworks:

  • Neglectedness: Many emerging technologies have minimal advocacy coverage
  • Scale: Potential impact involves trillions of animals
  • Tractability: Early-stage interventions are more feasible
  • Long-term Value: Preventing lock-in has persistent effects

Key Data Frameworks

The project draws on several important EA concepts:

Neglectedness, Tractability, Scale

Framework for identifying high-impact interventions in EA

Long-term Value = Significance × Persistence × Contingency

How big the impact is × How long it lasts × Whether acting now makes a unique difference

Suffering-Adjusted Days (SADs)

An attempt to measure total experienced suffering by combining duration and intensity

Rate of Subjective Experience

The idea that some species may experience time faster than humans

Capacity for Welfare

Estimating how much happiness or suffering a species is capable of

Problem Statement

Without early warning tools for animal farming lock-in, we risk:

  • Missing critical intervention windows
  • Allowing harmful systems to become entrenched
  • Creating permanent suffering for billions of animals
  • Reducing future options for animal welfare improvements

Solution Approach

LEWS addresses this problem by:

  • Providing quantitative risk assessment
  • Offering historical comparison to known lock-in events
  • Explicitly modeling uncertainty
  • Providing clear intervention recommendations
  • Being simple enough for non-experts to use

EA Alignment

The project is strongly aligned with effective altruism principles:

  • Neglected: Shrimp and insects have near-zero organization coverage
  • Scalable: Affects trillions of animals
  • Tractable: Early-stage interventions are cost-effective
  • Longtermist: Prevents permanent harmful equilibria

Project Timeline

As a hackathon project, LEWS was developed under tight constraints:

  • Pre-hackathon: Research and data preparation
  • Saturday: Team coordination and technical planning
  • Sunday (Hours 1-15): Implementation and integration
  • Sunday (Hours 15+): Testing, refinement, and presentation prep

Future Development

Potential enhancements beyond the hackathon MVP:

  • Full Bayesian model for uncertainty
  • Non-linear weighting systems
  • Cross-species calibration
  • Advanced ML for momentum estimation
  • Multi-technology comparison dashboard
  • Historical retrodiction features