Verification ledger

The journal’s integrity, audited in one place

Critical AI’s integrity claims are not asserted — they are re-checked, live, on every render. This ledger rolls up every piece: each evidence review’s aggregates re-derived from its codings, each synthesis’s citations re-checked against the reviewed corpus, and each per-paper critique validated against its access-basis severity cap. Machine-readable at /critique/integrity.json.

✓ Journal verifiedAll 14 published pieces pass their integrity checks, with 0 fabricated citations across the journal.

published pieces

0/0

reviews verified

14/14

critiques attested in-app

0/0

synthesis citations resolved

verbatim spans

fabricated citations

Every piece, checked

Type	Piece	Status	What was checked
critique	AI meets politics: Examining the effects of different targeting strategies across 15 countries	✓ verified	4 in-app checks pass (spans in app)
critique	Artificial Collusion: Examining Supracompetitive Pricing by Q-Learning Algorithms	✓ verified	4 in-app checks pass (spans in app)
critique	Generative AI at Work	✓ verified	4 in-app checks pass (checked sources)
critique	Can ChatGPT Kill User-Generated Q&A Platforms?	✓ verified	4 in-app checks pass (spans in app)
critique	Scaffolding Human–AI Collaboration: A Field Experiment on Behavioral Protocols and Cognitive Reframing	✓ verified	4 in-app checks pass (spans in app)
critique	From rule of law to rule of algorithm: Generative Artificial Intelligence's threat to democracy	✓ verified	4 in-app checks pass (spans in app)
critique	Generative AI, propaganda, and digital authoritarianism: Comparative insights from six democratically weakened countries	✓ verified	4 in-app checks pass (spans in app)
critique	Made With AI: Consumer Engagement with Social Media Containing AI Disclosures	✓ verified	4 in-app checks pass (spans in app)
critique	The Impact of AI on Developer Productivity: Evidence from GitHub Copilot	✓ verified	4 in-app checks pass (spans in app)
critique	Refusal as silence: Gendered disparities in Vision-Language Model responses	✓ verified	4 in-app checks pass (spans in app)
critique	The Cybernetic Teammate: A Field Experiment on Generative AI and Teamwork	✓ verified	4 in-app checks pass (spans in app)
critique	The politics of artificial intelligence alignment: Public reactions to AI moderation in the case of Google’s Gemini	✓ verified	4 in-app checks pass (spans in app)
critique	Unraveling Generative AI from a Human Intelligence Perspective: A Battery of Experiments	✓ verified	4 in-app checks pass (spans in app)
critique	Generative AI Adoption and Organizational Productivity: Evidence from 500 Firms	✓ verified	3 in-app checks pass (illustrative)

Critique calibration

Beyond integrity, the journal measures how close its critiques are to the published human-expert standard. Of 13 real critiques scored against the 39-critique benchmark corpus, 4 are calibrated (mean alignment 0.69, threshold 0.80). Abstract-only critiques score lower by design — the calibration page shows the access-basis gradient. This is a quality signal, not an integrity gate.

Ingestion provenance

The content set is ingested from the AGI Social Scientist by an auto-discovering pipeline: every AGISS review in the outputs is picked up automatically, so the journal tracks the engine without per-review code changes. Last ingested 2026-06-14 — 5 reviews + 1synthesis auto-discovered. Each review’s AGISS report hash anchors its offline re-derivation.