Feargate
feargate.org- Category
- Education
- Rank
- No. 138Apps board
Previous survey · No. 114 ·
- Type
- APP
- Date
About
An archive of public claims and warnings about AI with dates, links and the archive’s assessments. Readers can filter by speaker or subject and open the cited material.
At a glance
Feargate is an opinionated archive of AI predictions and risk warnings, pairing quotations with deadlines, subsequent evidence and verdicts. It is worth a look for finding the paper trail behind a particular statement. Treat its verdict as an argument to examine: the distinction between a warning and a prediction can change whether a deadline makes sense.
Good fit
Readers checking what was said, when it was said and which later reports bear on it. A claim page puts the quotation, deadline, evidence chronology and source links together. The source-image viewer also provides surrounding text, which can reveal qualifications lost in a short headline. That organization is useful even when you disagree with the verdict.
Consider something else when
You want a neutral census of AI forecasts or a statistical measure of a person's accuracy. The Hall of Shame assigns three points to a false claim and two to an overhyped one within the archive's selected set. Those totals depend on editorial selection and classification, without establishing a complete denominator of everything the person predicted.
Start here
Choose a subject in the archive, then open a claim and compare The original claim with The reality check and Follow the evidence. In the browser trial, Misinfo & elections returned six matching claims. Inspect the original source capture opened a screenshot viewer. Before repeating a verdict, check the quoted wording and follow the linked evidence for its scope and caveats.
Cost and requirements
The tested public browsing flow required no account, payment, installation or file upload. Claim pages have individual URLs you can cite. Original reports, archived pages and other evidence are hosted by separate sites, so access and availability depend on those publishers too. No private document or model API key is needed for the tested browsing workflow.
Limits to know
The method assigns five years when a prediction has no date and can label warnings about unrealized harms false. That rule needs scrutiny when the original statement expresses a possibility rather than promising an outcome. In the inspected election file, the original quotation expresses concern, while the archive presents a false verdict. A cited platform report describes limited observed impact alongside mitigation efforts. This review tested navigation and inspected one case; it does not certify the archive's verdicts.
Sources checked 2026-11-05. Source-based assessment.
Why it made the leaderboard
The source links help you distinguish a dated prediction from a broader warning and check how the archive reached its assessment.
Tags
Media
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.