How we track
AI layoffs.
Every entry in the register must satisfy the same three questions a reasonable reader would ask: what happened? who connected it to AI? how strong is that link? If any answer is missing or too vague, we don't publish.
Inclusion rules
An event is publishable only when all of the following are true:
- → A concrete employer-level workforce action: layoff, hiring freeze, attrition plan, restructuring, or role elimination.
- → A specific company attached to the event.
- → At least one public source URL.
- → A verbatim quote connecting the action to AI, automation, or AI-driven efficiency — or a clear evidence note explaining why the quote is unavailable.
- → A summary that can be written without overstating causality.
We never publish macro trend pieces, generic "AI will replace jobs" commentary, anonymous community claims without public corroboration, or layoffs where AI attribution is only an editor's guess.
Attribution types
Every event is labelled with the most precise attribution type we can defend:
direct_replacement Company/source says AI directly replaces people or tasks formerly done by people.
productivity_consolidation Company says AI increases output enough to need fewer workers.
restructuring_for_ai Company cuts roles while repositioning around AI products or AI investment.
automation_indirect Automation is one cause among several, but not the sole mechanism.
cited_but_unclear AI is mentioned in connection with the event, but the mechanism is weak or underspecified.
disputed Attribution is publicly challenged, contradicted, or materially unclear.
Attribution strength
Editors choose the lowest defensible strength, not the most exciting one.
explicit Source directly says AI/automation caused or will cause fewer roles.
implied Source strongly links AI efficiency/restructuring to cuts, but not in one direct statement.
speculative AI link is presented by a credible source but remains inferential or weak.
Confidence tiers
Confidence grades the quality of AI attribution — not the size of the event.
A first-party source or attributable executive quote exists and explicitly or strongly implies the AI connection. Typical: CEO/CFO quote, press release, SEC filing, earnings transcript.
The AI link is strong but first-party evidence is incomplete, indirect, or partly mediated by reporting. Typical: credible outlet quoting an internal memo, first-party statement implying AI efficiency without stating direct replacement.
The event itself is real, but AI attribution is thin, indirect, or weakly sourced. Typical: third-party inference, single-source report with limited attribution detail.
The layoff is real but the AI explanation is contested, materially contradicted, or a strong initial claim was later softened, narrowed, or denied.
Source hierarchy
We weight sources in this order:
- 1. First-party filings, transcripts, memos, IR/newsroom posts
- 2. Direct executive interviews in credible outlets
- 3. High-quality reported coverage
- 4. Aggregator / corroboration sources
- 5. Community / manual-only sources
Layoffs.fyi, TrueUp, WARN, Reddit, Blind, Fishbowl, and LinkedIn may support research but do not by themselves prove AI attribution.
The quote rule
The quote is the trust anchor. Every entry carries a verbatim quote connecting the workforce action to AI. We don't clean up wording beyond normal punctuation trimming.
If no verbatim quote is available, an event may still publish only if sourcing is otherwise solid and the evidence notes explicitly say why the quote is unavailable. We never publish based on unverified community claims.
When uncertain: publish later rather than sooner, choose lower confidence, choose weaker attribution strength, explain ambiguity in evidence notes, and prefer omission over overclaiming. The full editorial rules are version-controlled alongside the codebase.