Bullshit Detector

Will it have real-time capacity; will it work on text or speech or both?

We'll try to make it fully automated given some text initially. We may extend it to speech, at which point we would investigate real-time speech to text technology.

Will it be in app or website form?

Initially only in website format, and while we won’t be developing an app frontend initially there’s nothing to preclude that happening in the future.

Can people upload their own content or will it work using prescribed sources?

The detector will be a tool that people can freely apply to new textual content. Whether it applies to new domains of claims is unclear at this point, though I hope the knowledge base of apparently contradictory claims will be extensible.

Will it have any claims to objectivity or will it depend on what it is programmed to detect as bullshit?

The detector will rely on a manually-collected knowledge base of contradictory claims, so cannot claim objectivity.

How does it differ from other Australian fact checkers?

We’ve chosen a project that has little to do with traditional fact checking, and more relates to issues of consistency or integrity checking.

Will it be limited to GBR content, and if so why?

It should not be limited in capacity, from the perspective of the technology, but we’ll use GBR for a Proof of Concept application.