AI Lab Cope Tracker

The labs publish their own data. We score how hard they cope about what it means. Usage telemetry, labour research, public statements — all analysed through the Discontinuity Thesis.

Tracking: OpenAI Signals • Anthropic Labour Research • Google DeepMind • Meta AI

Lab Cope Leaderboard

Ranked by overall cope score. Data transparency + framing honesty = how much they try to hide displacement evidence.

Meta AI

Open-source framing as democratisation while cutting workforce

Data Transparency: 5/100 Framing Honesty: 20/100
82
TERMINAL COPIUM

Google DeepMind

Claims AI is a 'productivity multiplier', avoids displacement metrics

Data Transparency: 15/100 Framing Honesty: 30/100
78
HEAVY COPE

OpenAI

Published displacement data, framed it as 'adoption'

Data Transparency: 95/100 Framing Honesty: 25/100
72
HEAVY COPE

Anthropic

Found 75% programmer exposure, concluded 'no systematic unemployment'

Data Transparency: 85/100 Framing Honesty: 40/100
58
MODERATE
46%
OpenAI: Work = Doing
ChatGPT performs work, not just answers
75%
Anthropic: Programmer Exposure
Computer programmers most exposed to Claude
165
O*NET Activities Mapped
Job tasks measured by OpenAI Signals
4
Labs Tracked
OpenAI, Anthropic, Google, Meta

Lab Data Deep Dives

OpenAI Signals

21 months of ChatGPT usage data. Topics, demographics, work patterns, 165 O*NET job activities, geographic exposure. The most comprehensive displacement dataset ever published.

Anthropic Labour Research

Claude usage data mapped to occupational exposure. 75% programmer exposure, slowing younger-worker hiring. Published March 2026.

Google DeepMind

AlphaFold, Gemini, Veo, Concordia — and a research primer on AI in organisations. Extraordinary science framing. No displacement metrics. Published May 2026.

Geographic Exposure

118 countries and 50 US states ranked by per-capita AI usage. Knowledge economy capitals hit first: Singapore, Israel, Netherlands lead globally.

Cope Scores

Public statements about AI and jobs, scored for cope. The Volume Illusion, the Augmentation Myth, and other classic deflection patterns catalogued.

OpenAI: Work Messages — Ask vs Do vs Express

The “doing” category = ChatGPT performing actual work tasks.

CopeCheck Analysis

Nearly half of all work-related ChatGPT messages are “doing” — the tool has become the worker. This isn’t augmentation. It’s substitution, measured in OpenAI’s own telemetry.

The Cope Report

A weekly digest of AI displacement cope, scored by the Oracle.
Top stories, new verdicts, and fresh data.

Subscribe Free

Weekly. No spam. Unsubscribe anytime. Powered by beehiiv.