Why this publication exists
AI raises questions about work, business capability and who benefits. This publication gives those questions a stable home, with clear current answers and a visible way to revise them. I write in a personal capacity. Government receipt of a submission does not imply endorsement, and this site does not represent a university or receiving body.
Positions and open questions
The established positions come from my personal policy submissions. They support an Australian-controlled inference option, international partnerships and business choice; substantial justified public investment; additional renewable supply; proportionate water requirements; timely approvals; and practical SME capability. They oppose general AI consumption subsidies and support human accountability, staff participation and worthwhile work.
Questions about junior career routes, household affordability, full Australian net benefits and education remain open. Their articles provide provisional evidence assessments. I haven’t reached settled policy positions on those questions.
Connection to the research
My PhD examines how Australian SME leadership teams interpret AI and relate those interpretations to narrated resource commitments, routines and strategic responses. Its main interview study is inductive. This site does not publish participant material, private reflections, unpublished manuscripts or findings from that ongoing study. Research questions are not reported as results.
How evidence is treated
Each substantial article states its question, scope, current answer, confidence, evidence, limits, implications and what would change the answer. Experiments, observational associations, surveys, task-exposure assessments, modelled scenarios and forecasts are labelled separately. Confidence describes the strength and transferability of the stated answer; it is not a numerical probability unless a source supplies one.
Sources are checked through the date shown, rather than continuously. A newer tool can make an older experiment less transferable without making the original result false. The source register identifies observation windows and limitations. This is a bounded synthesis, not an exhaustive systematic review.
AI assistance and responsibility
Codex assisted with source discovery, drafting, implementation and checks for this site. AI output is not independent corroboration. My submitted texts record my established positions. I distinguish those from my current reading of open questions. Source checking and editorial review do not guarantee that every error has been found. I remain responsible for the positions published here.
This disclosure concerns this publication. It does not retroactively alter the language of the historical submissions.
How revisions work
The initial edition is v1, first published on 18/10/2026 in Adelaide. Each article has a current address and a preserved dated edition. Later editions must add a change record explaining the previous answer, new answer, evidence, confidence and practical consequences. A correction is dated when made; it is not backdated to the original submission.
I assess new evidence, prepare a revision and change record, and have the facts reviewed before approving a substantive change in my public position. Discovery does not itself authorise a public revision. The existing evidence watch is not a publisher, and no automatic publishing schedule is enabled.
Submission archive
Four completed personal consultation responses are preserved as the author’s copies. Submission, receipt, acceptance and recipient publication are distinct statuses. No recipient-hosted public URL for an individual submission has been verified. Official links identify the consultation, not a government publication of my text.
Only the public policy bodies are included. Private receipts, contact details, research files and the prepared unsent SA energy response are withheld. A submission’s availability or status is not treated as permission to republish third-party work. The archived bodies retain their original wording; current commentary belongs in separate articles.
Privacy
Public reader comparisons keep answers only in browser memory and never transmit them. There are no analytics, advertising trackers or reader accounts. The optional Ask Alex form sends a question and, only with explicit consent, an email address into a private review queue. The form explains retention and one-time notification limits. Owner review uses a separate protected session and private storage. Vercel delivers the pages and may process ordinary request data under its own service policies. Following a source link takes you to another organisation’s site, with its own privacy practices.