Navigating the AI Universe: AI in the Interests of Australia

By Steve Davies in partnership with Claude AI

On 15 July, Prime Minister Anthony Albanese addressed the University of Sydney on the government’s approach to AI: an Office of AI, a National Cabinet discussion this month, legislation flagged for early next year, and a “light-touch” framework built on “flexibility” rather than comprehensive rules. He invoked Medicare, the minimum wage, and women’s suffrage as the lineage this framework belongs to.

Applying the Deep Truth Persona – an analytical method built on Professor Albert Bandura’s mechanisms of moral disengagement – to the full text of that speech surfaces a pattern that doesn’t show up in any single sentence, only in the architecture of the whole. Six mechanisms register at intensity 4 or above, five of them at the maximum. Euphemistic labelling turns a deferred-regulation choice into a procedural virtue. Diffusion of responsibility spreads accountability across eight-plus ministerial portfolios without naming a single accountable office for any commitment. And in a speech the Prime Minister himself says concerns “all of us,” not one worker, artist, or affected-community voice appears anywhere in the text. Under high political context, that pattern escalates the finding to Red – not a verdict on the Prime Minister’s intent, but a finding about what the announcement makes testable, and what it doesn’t, yet.

That analysis is available in full elsewhere. This piece is about something that happened while doing it.

Partway through applying Deep Truth to the speech, working through it in partnership with Claude AI, something else became evident – something the mechanism-by-mechanism analysis, thorough as it is, doesn’t fully capture on its own. The question underneath the question. Not just: what does this speech say, and what is it doing? But: what is actually happening here, at a scale bigger than one speech, one policy, one government?

In a very real and practical sense, what is rapidly emerging – not someday, not hypothetically, but now, in the months either side of that speech – is an AI Universe. Not a new tool added to an old world, the way a calculator or a search engine was a tool. Something closer to a new domain of existence: a scope and speed of capability touching employment, creativity, governance, defence, and education all at once, reshaping how institutions explain themselves to the people they serve. No single person, government, or company built this universe on purpose. None can navigate it alone.

That’s the piece the PM’s speech analysis, on its own, couldn’t fully surface – because a mechanism-by-mechanism reading of one document, however rigorous, is necessarily a close-up. It is Aristotle’s old tenet playing out in real time: the whole is greater than the sum of its parts. Six mechanisms, each named, rated, and evidenced, are the parts. Step back from them, and a different thing comes into view – not a bigger version of the same finding, but a different kind of finding altogether: institutions everywhere, not just Australia’s, are being asked to govern a domain that didn’t exist in its current form five years ago, using language and processes built for a world of tools, not universes.

That’s why the partnership between AI and human judgement isn’t an optional extra to this analysis – it’s what made the analysis possible, and what revealed its own limits at the same time. AI brought the processing scale to hold Bandura’s eight mechanisms against thousands of words of transcript, consistently, without fatigue or selective attention. The human half of that partnership brought the thing no mechanism-tagging can generate on its own: the recognition that a well-executed analysis of one speech was, in the end, standing inside something much bigger than the speech itself.

Two pieces of work sit behind this one: a full Deep Truth analysis of the Prime Minister’s speech, and a separate exploration of what the AI Universe actually is and why the partnership needs a moral compass to navigate it. Neither needs this article to explain it. Read side by side, they make the same case from two directions – one from close up, one from a distance – and the case is stronger for not being told twice.

Deep Truth Persona V5.2 is available for public use under Creative Commons.

A note on the byline: the Deep Truth Persona, its underlying method, and four years of development are Steve Davies’ own work. “In partnership with Claude AI” refers to the drafting of this piece, not the authorship of the framework itself.

See also

Navigating the AI Universe

Steve Davies is a retired public servant. His expertise is in the areas of organisational research and people development. He’s always been attracted to forward looking work. He’s a vocal critic of destructive, cruel and backwards looking behaviours and practices.

Over the years he’s spoken in depth with whistleblowers and advocated the use of technology (including social media tech) to empower people to do great things together.

His thinking and work have been heavily influenced by such great thinkers and researchers as Shoshana Zuboff, Albert Bandura and Peter Senge for decades.