DataMPowered research explores how institutions can adopt AI without relocating knowledge, authority, judgement and accountability into models or ungoverned agent use.
The articles here form a connected body of thought: institutional capability, Operational Intelligence, human judgement, authority, accountability, AI governance, productivity, public-sector transformation, and the reliability and assurance questions that follow when models participate in work.
Publications may appear as essays, perspectives, research papers or briefs. The form varies; the focus remains how institutions preserve accountable judgement while making capability discoverable and operational.
That research informs ifCEM, DataMPowered's operational platform for governed professional work.
Read the founder's perspective on Operational Intelligence and accountable judgement inside real organisations.
Cloud, AI and the accounting question inside Australia's productivity puzzle
Cloud computing shifted productive capability from owned capital toward purchased services. Generative AI may move cognition the same way. This essay asks what that structural shift means for Australia's productivity puzzle—not whether national accounts simply forgot to count SaaS.
AI does not need to seize control. We may simply keep accelerating the systems we increasingly depend on to decide what comes next.
Capability can outrun the institutions meant to understand, judge and constrain it. An essay on rational dependence, the intelligence–authority relationship, and why better machine judgement does not settle who may decide.
What organisations need to know before 10 December 2026 and how ifCEM could help
From 10 December 2026, APP entities may need to disclose in their privacy policies where computer programs participate in consequential decisions. The disclosure is the visible part. The harder work is operational.
Five areas of inquiry connect the corpus. Each pillar groups articles that examine a distinct institutional question—not a chronological archive.
Browse every research article in one place.
How institutions supervise AI-assisted work, manage bias and reliability risk, and keep evidence, accountability and human oversight in view when models contribute to consequential decisions.
Why enterprise AI adoption depends on identity, authority, policy, knowledge and execution remaining institutionally grounded—not relocated into models, agents or parallel AI institutions.
Where human authority, legitimacy and accountable judgement must remain when AI capability grows—including questions of superintelligence, alignment and who should decide.
What productivity means when machine cognition becomes cheap: unfinished digital transformation, institutional capability, judgement and the gap between AI consumption and durable institutional gain.
How AI changes demand, access and service delivery in the public sector—and why removing administrative friction may require intent-centred government rather than more portals and forms.
Cloud, AI and the accounting question inside Australia's productivity puzzle
Cloud computing shifted productive capability from owned capital toward purchased services. Generative AI may move cognition the same way. This essay asks what that structural shift means for Australia's productivity puzzle—not whether national accounts simply forgot to count SaaS.
AI does not need to seize control. We may simply keep accelerating the systems we increasingly depend on to decide what comes next.
Capability can outrun the institutions meant to understand, judge and constrain it. An essay on rational dependence, the intelligence–authority relationship, and why better machine judgement does not settle who may decide.
What organisations need to know before 10 December 2026 and how ifCEM could help
From 10 December 2026, APP entities may need to disclose in their privacy policies where computer programs participate in consequential decisions. The disclosure is the visible part. The harder work is operational.
The contest over models is real. A second contest, over where capability and judgement live, is still open.
Stronger AI need not absorb the capability of people and institutions. This essay tests whether a second trajectory is technically credible: models that help people reach capability which stays local, authoritative and replaceable.
Technical debt, workarounds and legacy systems have accumulated for decades. AI is beginning to remove the human friction that made those environments difficult to understand, combine and traverse.
The vulnerability is not new. The cost of discovering, understanding and combining it is changing. Frontier AI reduces the cognitive friction that once limited how quickly institutional weakness could be found and chained together.
Task capability is not role substitutability. Reliability is not authority. And keeping the outputs is not keeping the capability.
AI can increasingly perform the observable tasks of a role while the institution quietly loses the capability the role carried. The more useful question is not which jobs AI will take, but what the institution must still be able to do for itself.
What Salesforce Koa, the pursuit of superintelligence and a warning from Sam Altman reveal about institutional capability
Salesforce Koa shows how accumulated enterprise operating knowledge can become specialised model capability. The institutional question is what should enter the model, and what must remain independently authoritative.
The national return on AI investment depends on whether compute and models translate into productive, sovereign and accountable capability.
Australia is investing in AI infrastructure and training. The larger question is how compute and models translate into productive, sovereign and accountable Australian capability.
Establish what can be established. Use inference for what genuinely needs to be inferred.
AI can make existing organisational capability easier to discover and use. But we may be taking a different path, reconstructing knowledge, rules, authority and execution inside increasingly capable AI systems.
When cognition becomes cheap, what remains scarce?
When machine cognition becomes cheap, productivity depends on more than consuming more AI. This essay examines Australia's productivity puzzle, the economics of cheap cognition, institutional capability, judgement and the unfinished digital transformation.
The uncomfortable gap between what AI can do and what institutions can responsibly allow it to do
AI can look extraordinary when raw capability is enough. Inside an institution, capability must also operate within identity, authority, policy, accountability and judgement. What if that explains why AI can be easier to weaponise than to institutionalise?
The distinction matters, because intelligence is only one part of the problem.
What exactly are we calling superintelligence, and why do we assume that becoming more intelligent should naturally mean acquiring more agency, authority and power? The distinction matters, because intelligence is only one part of the problem.
Digitisation taught humans to translate intent into systems. AI is beginning to make that translation happen in real time. That may force us to rethink not only jobs, but applications, management, education, organisational knowledge and the skills we choose to build next.
Digitisation taught people to translate human intent into systems, applications and specialised processes. AI is beginning to collapse that translation layer. What does that mean for jobs, skills, management, education and institutional capability?
The opportunity may not be to rebuild organisations around AI, but to make institutional capability easier for people to discover, assemble and use.
We may be treating AI as another enterprise transformation. The opportunity may be a language interface into capability institutions already have.
Why digital government may need to move from citizen navigation to intent-centred service delivery
AI may not simply create demand for government services. It may expose demand that administrative friction previously concealed, and that changes the interface between citizen and institution.
Why AI Could Reconnect Organisations With Their Own Logic
For decades, organisations have used technology to solve business problems by creating new layers of systems, applications, platforms, workflows, databases, dashboards, and vendor-managed environments. Each layer promised efficiency. And in many ways, each delivered. But over time, those same layers also moved organisational knowledge further away from the people who needed to understand, question, adapt, and act on it.
Regaining Organisational Control in the Age of Outcome-Driven AI
A response to the tension between messy organisational reality and outcome-driven AI, and why the next phase of adoption depends on an operational layer that restores visibility, governance, and accountability.
Designing Governed Systems for Human-Centric Intelligence
A comprehensive analysis of how bias permeates AI systems at every stage, from data collection through human interaction, and governance strategies for building accountable, human-centric AI.
Causes, Risks, and Architectural Strategies for Prevention
A systems-first examination of AI hallucinations, exploring the architectural and organizational strategies to mitigate hallucination risk in production environments.
How institutions supervise AI-assisted work, manage bias and reliability risk, and keep evidence, accountability and human oversight in view when models contribute to consequential decisions.
What organisations need to know before 10 December 2026 and how ifCEM could help
From 10 December 2026, APP entities may need to disclose in their privacy policies where computer programs participate in consequential decisions. The disclosure is the visible part. The harder work is operational.
Designing Governed Systems for Human-Centric Intelligence
A comprehensive analysis of how bias permeates AI systems at every stage, from data collection through human interaction, and governance strategies for building accountable, human-centric AI.
Causes, Risks, and Architectural Strategies for Prevention
A systems-first examination of AI hallucinations, exploring the architectural and organizational strategies to mitigate hallucination risk in production environments.
Why enterprise AI adoption depends on identity, authority, policy, knowledge and execution remaining institutionally grounded—not relocated into models, agents or parallel AI institutions.
The contest over models is real. A second contest, over where capability and judgement live, is still open.
Stronger AI need not absorb the capability of people and institutions. This essay tests whether a second trajectory is technically credible: models that help people reach capability which stays local, authoritative and replaceable.
Technical debt, workarounds and legacy systems have accumulated for decades. AI is beginning to remove the human friction that made those environments difficult to understand, combine and traverse.
The vulnerability is not new. The cost of discovering, understanding and combining it is changing. Frontier AI reduces the cognitive friction that once limited how quickly institutional weakness could be found and chained together.
Task capability is not role substitutability. Reliability is not authority. And keeping the outputs is not keeping the capability.
AI can increasingly perform the observable tasks of a role while the institution quietly loses the capability the role carried. The more useful question is not which jobs AI will take, but what the institution must still be able to do for itself.
What Salesforce Koa, the pursuit of superintelligence and a warning from Sam Altman reveal about institutional capability
Salesforce Koa shows how accumulated enterprise operating knowledge can become specialised model capability. The institutional question is what should enter the model, and what must remain independently authoritative.
Establish what can be established. Use inference for what genuinely needs to be inferred.
AI can make existing organisational capability easier to discover and use. But we may be taking a different path, reconstructing knowledge, rules, authority and execution inside increasingly capable AI systems.
The uncomfortable gap between what AI can do and what institutions can responsibly allow it to do
AI can look extraordinary when raw capability is enough. Inside an institution, capability must also operate within identity, authority, policy, accountability and judgement. What if that explains why AI can be easier to weaponise than to institutionalise?
The opportunity may not be to rebuild organisations around AI, but to make institutional capability easier for people to discover, assemble and use.
We may be treating AI as another enterprise transformation. The opportunity may be a language interface into capability institutions already have.
Why AI Could Reconnect Organisations With Their Own Logic
For decades, organisations have used technology to solve business problems by creating new layers of systems, applications, platforms, workflows, databases, dashboards, and vendor-managed environments. Each layer promised efficiency. And in many ways, each delivered. But over time, those same layers also moved organisational knowledge further away from the people who needed to understand, question, adapt, and act on it.
Regaining Organisational Control in the Age of Outcome-Driven AI
A response to the tension between messy organisational reality and outcome-driven AI, and why the next phase of adoption depends on an operational layer that restores visibility, governance, and accountability.
What productivity means when machine cognition becomes cheap: unfinished digital transformation, institutional capability, judgement and the gap between AI consumption and durable institutional gain.
Cloud, AI and the accounting question inside Australia's productivity puzzle
Cloud computing shifted productive capability from owned capital toward purchased services. Generative AI may move cognition the same way. This essay asks what that structural shift means for Australia's productivity puzzle—not whether national accounts simply forgot to count SaaS.
The national return on AI investment depends on whether compute and models translate into productive, sovereign and accountable capability.
Australia is investing in AI infrastructure and training. The larger question is how compute and models translate into productive, sovereign and accountable Australian capability.
When cognition becomes cheap, what remains scarce?
When machine cognition becomes cheap, productivity depends on more than consuming more AI. This essay examines Australia's productivity puzzle, the economics of cheap cognition, institutional capability, judgement and the unfinished digital transformation.
Digitisation taught humans to translate intent into systems. AI is beginning to make that translation happen in real time. That may force us to rethink not only jobs, but applications, management, education, organisational knowledge and the skills we choose to build next.
Digitisation taught people to translate human intent into systems, applications and specialised processes. AI is beginning to collapse that translation layer. What does that mean for jobs, skills, management, education and institutional capability?
How AI changes demand, access and service delivery in the public sector—and why removing administrative friction may require intent-centred government rather than more portals and forms.
Why digital government may need to move from citizen navigation to intent-centred service delivery
AI may not simply create demand for government services. It may expose demand that administrative friction previously concealed, and that changes the interface between citizen and institution.
Our research informs how ifCEM assembles context, authority and capability around real work, and records material contributions, including models where they assist.
Explore ifCEM →