Who publishes this work
Opening Research is published and funded by Opening, a commercial career-technology company. It is not a separate university, government agency or independently governed nonprofit. Opening benefits commercially from being trusted on labor markets; that interest is relevant to how readers evaluate its work.
The editorial position supports workers’ pay, safety, voice and right to organize. It applies the same evidentiary standards to claims from companies, governments, unions and technology vendors. A desired political or commercial outcome is not evidence.
How this launch was produced
The launch reporting was prepared with AI assistance for research, drafting, source comparison and technical development. The byline “Opening Research” identifies the publisher, not an invented individual journalist. These articles have not undergone academic peer review or an independent human editorial review. They synthesize published evidence; no original interviews are claimed.
Readers can inspect linked sources and the stated limits of each article. AI assistance does not remove Opening’s responsibility for accuracy. Claims that could not be supported were excluded or held for further reporting.
Evidence and attribution
Primary records, official statistics and original research take precedence. Published papers and working papers are distinguished, and the cited version matters. A finding keeps its population, period and outcome. Sources appear next to the passages they support and in the article’s reference list.
Forecasts, task exposure, adoption, measured outcomes and causal effects are different evidence categories. A job advertisement is not a hire. An association with AI exposure is not proof of an AI-caused layoff. A company’s explanation for a workforce change is attributed to the company until independently established.
Published source definitions and limitations are public. Private career records and Opening’s proprietary collection and scoring systems are not part of the public dataset. A private sample will not be presented as a national census.
Data freshness and reproducibility
The data desk displays official BLS estimates and archives complete retrieval vintages. The observation month and retrieval timestamp are distinct. The source is checked daily; its publication schedule determines when new estimates exist. The latest complete vintage remains available if a refresh fails, with a warning when retrieval is overdue.
Downloads identify the source, series, unit, adjustment and vintage. Gaps are not replaced with zeros. Revisions are stored in new vintages. Articles retain their evidence date and do not silently acquire a new conclusion when the dashboard refreshes.
Corrections and updates
Material corrections identify the affected claim, the change and the correction date on the article. Routine copy edits do not reset the evidence date. A changed source estimate is a data revision; an error in Opening’s transcription or interpretation is a correction. Both should remain visible when they affect a conclusion.
Send a disputed passage, supporting source and relevant date to the research corrections inbox. This is an ordinary email address, not a secure confidential tip channel. Do not send sensitive employment or organizing records without arranging an appropriate channel first.
Launch correction record — September 9, 2026. The earlier ghost-jobs article has been replaced by a measurement guide that removes unsupported prevalence and intent claims. Other earlier commentary remains identified as under source review. The pay-disclosure sample tracker remains unpublished while its missing-data interpretation is reviewed.
Citing and reusing the work
Link to the article, name Opening Research and include its publication date. For an external study’s estimate, cite the original study. For the dashboard, credit BLS as producer and Opening Research for the presentation or analysis, and retain the downloaded retrieval vintage.
The CSV and JSON downloads contain public federal statistical observations. Source terms govern source material; access through Opening does not change those terms. Contact the research desk for republication of full Opening articles or custom analysis. The RSS feed provides article titles, descriptions and links.
Reporting commitments
Future investigations will require primary documents, a clear distinction between allegation and established fact, an opportunity for relevant subjects to respond, and disclosure of material conflicts. Interviews, partnerships, grants and institutional independence will be described only when they exist. Planned work is labeled as a reporting question rather than a completed finding.