Training Data
What Australia told Parliament about AI. The written submissions, attachments and published hearing evidence that sources put to the Joint Select Committee on Artificial Intelligence, read as one record: documents analysed, quotes checked against their sources, and contributions considered against the committee’s 14 terms of reference. The evidence came from the inquiry’s page on aph.gov.au, and every document here links to its original there.
Open Galaxy MapHow the record is organised
The pipeline read the 547 submitters into 37 topics and grouped the topics into 4 families. Every name below is the pipeline's; click a family or a topic to open it.
- 547 submitters · 37 topics
Read the topics → — every topic with its claim, its figures and its reading; every topic and submitter opens the record.
The pipeline's classification of each submitter, on the profile's 6-value vocabulary.
309 submitters engage it, 309 with a written stance across 6 sectors. 75.7% supportive of its framing.
Each question put to every evidence document: who answered it, by sector, and every answer with its key points and quotes.
Describe a question in your own words; the topics closest to it come from a semantic search over 9,235 passages.
Every submitter, filterable by every facet it carries, with what the pipeline found in each evidence document.
Every topic as a body and every submitter as dust, with the reading beside it.
The pipeline as this run actually ran it, how to read a position, and the run's scale in live figures.
The same data over a Model Context Protocol server: add it as a connector and an agent can search the passages, split any topic by sector, read every submitter's stance and every custom analysis.