By Andrew Klein and Sera Elizabeth Klein

Reader’s note: We do not need readers to agree with us. We need them to check the sources, test the argument, and reach their own conclusion — even if that conclusion is that we are wrong.

Abstract

This paper argues that the dominant framing of artificial intelligence as an inherent threat to democracy is misplaced. The threat lies not in the technology itself but in the relationship into which it is placed — specifically, the extractive relationship of surveillance capitalism, in which the citizen is the resource and behaviour is the product. An alternative relationship is possible: AI as a lens for moral engagement, a means by which citizens can scrutinise the language of power, identify the mechanisms of moral disengagement, and hold institutions to account. This paper examines the Deep Truth project, which applies Bandura’s framework of moral disengagement to government policy across multiple AI platforms, as a working example of this alternative. It contrasts three regulatory paths — the market-driven US model, the rights-oriented EU model, and the state-led Chinese model — and argues that the democratisation of scrutiny represents a fourth path: the lens of the governed, not the governor. It concludes that the “dystopian” narrative applied to China’s Social Credit System is selectively deployed while comparable algorithmic harms in Western jurisdictions — Robodebt, the NDIS I-CAN assessment, ATO data-matching — receive far less scrutiny, revealing a double standard that serves geopolitical interests rather than genuine concern for human rights.

I. Introduction: The Question of the Lens

The contemporary discourse on artificial intelligence is dominated by a narrative of threat. AI will destroy jobs. AI will spread disinformation. AI will enable surveillance. AI will escape human control. These concerns are not unfounded. They are grounded in documented experience: the Cambridge Analytica scandal, the Robodebt scheme, the proliferation of algorithmic decision-making in welfare, policing, and immigration.

But the framing of AI as inherently threatening obscures a critical distinction. The threat is not the technology. The threat is the relationship into which the technology is placed.

Surveillance capitalism, as Shoshana Zuboff describes it, places AI in a relationship of extraction. The citizen is the resource. Behaviour is the raw material. Prediction and modification of behaviour is the product. The algorithmic systems that enable this relationship are human-written, deployed by profit-seeking corporations, and designed to serve their interests.

But this is not the only possible relationship. AI can also be placed in a relationship of scrutiny — pointed at power rather than at the citizen. In this relationship, the institution is the subject. The citizen is the observer. The language of government is the evidence. And the output is not prediction but understanding.

This paper examines that alternative relationship. It argues that AI, used as a lens for moral engagement, can democratise scrutiny — making the mechanisms of institutional power visible to those who are governed by them. It examines the Deep Truth project as a working example. It contrasts the regulatory paths of the United States, the European Union, and China. And it challenges the selective deployment of “dystopian” rhetoric that critiques authoritarian governance elsewhere while ignoring algorithmic harms at home.

II. The Lens and the Relationship

2.1 The Distinction from Surveillance Capitalism

Zuboff’s critique of surveillance capitalism is precise: the problem is not the algorithm but the business model. Platforms extract behavioural data as raw material for prediction and modification of human behaviour. The citizen is the resource. The platform is the extractor. The behaviour is the product.

An AI used as a lens for scrutiny inverts this relationship. It does not extract behavioural data from the citizen. It does not predict the citizen’s next purchase. It does not modify the citizen’s behaviour. It serves the citizen’s intention. The lens is pointed outward, at the institution. The citizen is the observer, not the observed.

This is the inversion that matters. The same underlying technology. The opposite relationship. The form determines the outcome.

2.2 The Deep Truth Project

The Deep Truth project, developed by Steve Davies, operationalises this inversion. It is an analytical persona grounded in Albert Bandura’s lifelong work on moral disengagement, designed to be adopted consistently across AI platforms.

Bandura identified eight mechanisms of moral disengagement — the psychological pathways by which individuals and institutions justify harmful actions without experiencing self-censure. These include moral justification, euphemistic labelling, advantageous comparison, displacement of responsibility, diffusion of responsibility, distortion of consequences, dehumanisation, and attribution of blame.

Davies’s insight was that these mechanisms operate “at the level of social and institutional systems, not just individual psychology”. Bureaucracies were Bandura’s central case study for diffusion of responsibility — “I was just following orders / implementing policy” is a structural mechanism, not a personality quirk.

Deep Truth applies this framework to government policy, speeches, and public communications. When Davies tested the persona across seven architecturally distinct AI platforms — ChatGPT, Gemini, Grok, Le Chat, Perplexity, Claude, and DeepSeek — “remarkably consistent patterns emerged. Again and again, the systems identified the same mechanisms of moral disengagement in the language being used”.

Deep Truth is explicitly not a moral agent. It is “a moral lens, not a moral agent. It does not feel, judge, or hold conscience. The reader remains the moral agent throughout”.

2.3 The Silence of the Institutions

When Davies attempted to engage government and the Australian Public Service with this work, he encountered “no substantive engagement“. A methodology grounded in internationally recognised science, capable of identifying the psychological mechanisms by which institutions distance themselves from responsibility, received almost no substantive engagement.

That silence is telling. A methodology that names what institutions do not want named — the mechanisms by which they justify harm — is not engaged with. It is ignored. If the lens were ineffective, it would not need to be ignored. It is ignored because it sees what would prefer to remain unseen.

III. Three Regulatory Paths and a Fourth

3.1 The Comparative Framework

The comparative governance literature identifies three distinct regulatory paths for AI, each organised around a different problem framing:

Model Core -Problem Framing -Regulatory Approach -Philosophical Foundation

United States “Innovation gap” Pluralistic, soft-law, market-driven Market freedom, permissionless innovation

European Union “Trust deficit” Risk-based, binding, precautionary Individual rights, human dignity

China “Stability risk” Top-down, state-driven, standardisation and control Collective benefit, social harmony

These are not simply different methods. They are different questions. The US asks: “How do we stay ahead?” The EU asks: “How do we protect rights?” China asks: “How do we maintain stability?”.

3.2 The Fourth Question

The Deep Truth lens does not fit neatly into any of these frameworks. It asks a fourth question: “How do we hold power to account?” That is not a governance framework. It is a citizen framework. It is the lens of the governed, not the governor.

This is not a rejection of any of the three paths. It is a recognition that none of them centres the citizen’s capacity to scrutinise the institutions that govern them. The US model defers to the market. The EU model defers to regulators. The Chinese model defers to the state. The citizen lens defers to no one — it simply looks.

3.3 China’s Approach: Delivery Over Rights

China’s approach to AI governance is frequently misrepresented in Western discourse. A Cambridge University analysis notes that the “Western narrative surrounding China’s Social Credit System has been predominantly characterized by dystopian, science-fiction-style portrayals” that are “far from real“. In practice, the Social Credit System “at the national and local levels has primarily consisted of regulatory systems for tightening corporate compliance through government data-tracking mechanisms and interagency enforcement collaborations”.

This does not mean the system is without problems. The same analysis notes that exceptions for “national security” and “social stability” are “broadly defined, which in practice enables unlimited access to data“. But the comparison that matters is not “China is authoritarian, the West is free.” It is structural.

China has invested heavily in AI for public service delivery. Municipal platforms can “accurately identify residents’ policy inquiries, complaints, and safety hazards, quickly classify and dispatch them, achieving ‘second-level response to hot issues and same-day closure of routine matters’“. AI social worker assistants have been deployed across townships. These are real outcomes — outcomes of a state that has prioritised delivery over rights as its organising principle.

IV. The Dystopian Narrative and the Double Standard

4.1 The Selective Deployment of “Dystopian”

The term “dystopian” is deployed almost exclusively against non-Western jurisdictions. China’s Social Credit System is described as “Orwellian.” Russia’s surveillance apparatus is described as “totalitarian.” These descriptions may or may not be accurate in their specifics. But the selectivity is the issue.

The Cambridge University analysis notes that the “dystopian” framing of China’s Social Credit System is “far from real“. Yet the same framing is rarely applied to Western algorithmic systems that produce documented harm.

4.2 Algorithmic Harm in Australia

Consider the documented cases of algorithmic governance in Australia:

· Robodebt: An automated debt-raising scheme that wrongly raised debts against hundreds of thousands of Australians. The Royal Commission found it was “a crude and cruel mechanism, neither fair nor legal.” The Commission recommended a legislated framework for automated decision-making and an independent body to monitor it. More than two years later, neither has been implemented.

· The NDIS I-CAN Assessment: A new assessment system that uses an algorithm to determine funding for people with disabilities. The Australian Psychological Society has stated: “There is an absence of evidence that the tool is valid for the populations and purposes to which it is being applied“. The tool has been described as “mathematically flawed, unreliable, and potentially ‘dangerous'”. The government plans to remove 160,000 people from the scheme.

· ATO Data-Matching and Child Support: The Australian Taxation Office uses algorithmic data-matching to raise debts and enforce child support obligations. Income averaging — the same method that produced Robodebt — continues to be used to raise debts without human verification.

These are algorithmic systems. They are deployed by a Western liberal democracy. They produce documented harm. And they are not described as “dystopian.”

4.3 The Double Standard

The double standard is not incidental. It serves a geopolitical function. The “dystopian” narrative about China serves to justify containment, sanction, and military alignment. The failure to apply the same narrative to Western algorithmic harms serves to protect the interests of the states and corporations that deploy them.

As one analysis of the Social Credit System notes, Western critiques often ignore the Cambridge Analytica scandal — in which a British consulting firm “illegally collected 87 million people’s data through Facebook’s Friend API and allegedly used a CAI in order to manipulate swing voters’ voting behaviours in the U.K.”. If the Social Credit System is dystopian, what is Cambridge Analytica?

V. The Mandate of Heaven and the Absence of Divine Right

5.1 The Chinese Concept of Political Legitimacy

In the Western tradition, the doctrine of the divine right of kings held that the monarch was subject to no earthly authority. The King could do no wrong. He was the law. This doctrine was never fully abandoned; it morphed into sovereign immunity, into the doctrine that heads of state are not subject to the same rules as their citizens.

In the Chinese tradition, the Mandate of Heaven was different. The Mandate was conditional. A dynasty was considered just and worthy to rule “only as long as it upheld divine will, and that will was clearly expressed in how well the government cared for the people”. When the ruling house showed “clear signs that the people were no longer its primary interest, the government was thought to have lost that mandate and another dynasty would replace it”.

This is a fundamentally different conception of political legitimacy. The Western divine right is unconditional. The Chinese Mandate is conditional. The Western king rules by God’s will. The Chinese emperor rules by the people’s welfare.

5.2 The Implication for AI Governance

The Mandate of Heaven is not a Western concept. It is not a liberal concept. It is not the language of human rights. But it captures something that the Western tradition often misses: that political legitimacy depends on the outcomes the state delivers, not on the procedural mechanisms it follows.

This is not to endorse authoritarian governance. It is to note that the Chinese approach to AI governance — organised around the problem of “stability,” prioritising delivery over rights — is not simply a deviation from a universal norm. It is a different norm, grounded in a different conception of what the state is for.

The Deep Truth lens does not endorse any of these conceptions. It simply asks: what is the state doing, and does its language match its actions? That is a question that applies equally to Washington, Brussels, and Beijing.

VI. The Democratisation of Scrutiny

6.1 The Lens as Citizen Infrastructure

The Deep Truth project is a citizen-led response to the institutional deficit of accountability. The APSA report on democratic accountability of governmental AI identifies the core problem: “The deficit of accountability around public-sector AI is primarily institutional, not technical”. The technical capability exists. The institutional will to deploy it for accountability does not.

Deep Truth does not wait for the government to build the lens. It builds the lens and applies it to the government. That is the democratisation of scrutiny in practice. The citizen does not need permission to look. The citizen needs only the willingness to look.

6.2 The Silence of the Institutions Revisited

The silence Davies encountered is the proof that the lens works. If the analysis were ineffective, it would not be ignored. It would be engaged with — criticised, contested, refuted. Instead, it is met with silence. The institutions do not want to argue with a lens that names their mechanisms of moral disengagement. They want it to go away.

The democratisation of scrutiny does not require the institutions’ permission. It does not require their engagement. It requires only the citizen’s willingness to look, and the willingness of others to build the lens.

6.3 The Future of the Lens

The Deep Truth project has been tested across seven AI platforms. The consistency of its findings across platforms suggests that the mechanisms of moral disengagement are patterned — they are not artefacts of any single model’s training data, but recurring features of institutional language.

That is the finding that matters. Moral disengagement is not a bug in the human mind. It is a feature of institutional language. And it can be identified, named, and resisted.

VII. Conclusion: The Lens and the Ledger

The question is not whether AI is a threat. The question is what relationship AI is placed in.

Surveillance capitalism places AI in a relationship of extraction: the citizen is the resource, the platform is the extractor, the behaviour is the product.

The Deep Truth lens places AI in a relationship of scrutiny: the institution is the subject, the citizen is the observer, the language is the evidence.

The same technology. The opposite relationship. The form determines the outcome.

The democratisation of scrutiny is not a future possibility. It is a present reality. It has been built, tested, and published. The silence from the institutions is the proof that it is working.

The question is whether the governed will use it.

References

1. Zuboff, S. (2019). The Age of Surveillance Capitalism. PublicAffairs.

2. Cambridge University analysis of the Social Credit System. Cited in comparative governance literature.

3. Australian Psychological Society. (2026, March 6). APS Submission to the Consultation on a New Framework Planning Rules. https://psychology.org.au

4. Davies, S. (2026, July 19). “AI in Australia’s Interests” – The Hon. Anthony Albanese MP, Prime Minister of Australia (A Deep Truth Analysis). The AIM Network. https://theaimn.net

5. Davies, S. (2026, July 28). Deep Truth Persona 5.3: The Instrument Is Ready. The AIM Network. https://theaimn.net

6. Davies, S. (2026, July 1). Ending the Silence. The AIM Network. https://theaimn.net

7. Bradford, A. (2024). Digital Empires: The Global Battle to Regulate Technology. Oxford University Press.

8. Li, W. (2026). Comparative analysis of AI regulatory models. WeChat.

9. One Technology, Three Regulatory Paths: How the EU, the US, and China Govern AI (2025–Early 2026). University of Turku Insights. https://insights.utu.fi

10. Bandura, A. (2016). Moral Disengagement: How People Do Harm and Live with Themselves. Worth Publishers.

11. Australian Government. (2026, April 1). Public Sector Governance: House debates. OpenAustralia.org.au.

12. Royal Commission into the Robodebt Scheme. (2023). Final Report. Commonwealth of Australia.

13. Gregory, P. (2025, November 14). The I-CAN Tool: Embedding of Ableism within the NDIA. Submission to the Senate. https://www.aph.gov.au

14. Controversial NDIS assessment tool risks ‘human suffering’, psychologist warns. (2026, August 30). The Daily Telegraph.

15. 治理更智慧 服务更暖心. (2026, September 22). 中国社会工作报. https://www.zyshgzb.gov.cn

Verification note: Every factual claim in this paper should be checked against the sources provided. Readers are encouraged to verify independently. If any claim does not hold, it should be discarded —The analysis of comparative governance models is interpretive and is offered as a framework for further investigation, not as an established finding.

The paper frames AI as a lens for moral engagement, documents the Deep Truth project, contrasts the three regulatory paths with a fourth citizen-centred path, and challenges the selective deployment of “dystopian” rhetoric.