A July 2026 report from the Tony Blair Institute for Global Change is drawing scrutiny from privacy advocates and policy analysts alike. The document, titled ‘Where Do I Start?’: How Governments Can Use AI and Data to Unlock Fiscal Space, explicitly calls on governments to deploy artificial intelligence alongside digital identity systems to “rank taxpayers” by financial risk and to target “discrepancies between taxpayer activity and declarations.” The report frames this as a fiscally responsible “dual-track approach” to public financial management — but the architecture it describes bears examination on its own terms.

What the Report Actually Proposes

Published on July 22, 2026, the TBI insight report outlines what it calls a dual-track approach to help governments “identify revenue loss and spending leakage.” The document identifies a set of AI-enabled capabilities it considers most relevant to this goal. Among them, two stand out in their scope:

The first is cross-system entity matching, described in the report as the ability to “link people, firms, accounts, tax IDs, suppliers and beneficiaries across fragmented databases.” The second is explainable risk scoring, defined as the capacity to “rank taxpayers, suppliers, refunds, payments or declarations by likely fiscal risk.”

Critically, the report notes that “many governments already hold enough usable data to start” this process. The data it references includes payroll files, payment records, procurement systems, customs declarations, tax returns, beneficiary registries, company registers, geographic information system data, and — notably — digital ID records.

The report’s “long-track” component strengthens what TBI describes as “the digital and fiscal infrastructure of the state,” encompassing “financial management, tax administration, payroll, procurement, customs, digital identity, payments, data standards and institutional controls.”

Digital ID as the Connective Layer

The Blair Institute has been an outspoken advocate for digital ID systems for several years. On its own website, TBI argues that digital ID systems “improve governance, facilitate greater inclusion, fuel economic growth and help governments achieve their core goals,” and that “far from enabling greater surveillance, digital IDs can actually make information more secure.”

TBI’s own analysis, referenced in a 2025 video address by Tony Blair, estimates that introducing digital ID in the UK “would save at least £2 billion a year” through “reducing losses to fraud, improving tax collection and better targeting of states abroad.” Blair has pointed to Estonia, Singapore, and India as models, noting that Estonia’s digital ID system allows citizens to file annual tax returns in an average of three minutes.

TBI’s website further describes digital ID as enabling citizens to “access government services through a single unique identifier” — a centralizing principle that the July 2026 fiscal report now explicitly connects to AI-driven financial risk assessment.

What the July 2026 report makes structurally apparent is the role digital ID plays as the connective tissue in this system: without a universal unique identifier tethered to an individual’s financial activity, cross-system entity matching and taxpayer risk scoring at the scale envisioned would not be technically feasible.

The Risk-Scoring Architecture: How It Works

The term “explainable risk scoring” is drawn from the field of machine learning, where it refers to AI models designed to produce outputs that human reviewers can interpret and audit. In the context TBI describes, the scoring system would ingest data from multiple government databases — tax returns, customs declarations, payment records, payroll files — and produce a ranked assessment of each taxpayer’s likelihood of fiscal non-compliance.

This is not, as TBI frames it, simply a modern audit tool. It is a system in which an algorithmic score, generated by correlating data points across a citizen’s financial life, determines the level of government scrutiny that individual receives. Those ranked as higher risk face greater examination of their declarations; those ranked lower face less.

The practical implication is that an individual’s relationship with their tax authority would be mediated, at least in part, by a score they did not generate, cannot directly inspect, and may have limited ability to contest — even if TBI’s emphasis on “explainability” is taken at face value.

Comparisons to Existing Scoring Frameworks

Observers have drawn comparisons between the architecture TBI proposes and elements of China’s social credit system, which Wikipedia describes as “a national credit rating and blacklist implemented by the government of the People’s Republic of China” designed to “track businesses, individuals, and government institutions and enhance their perceived trustworthiness.” China’s system uses varying degrees of whitelisting and blacklisting and incorporates big data analytics drawn from financial transactions and other behavioral data.

It is worth noting distinctions as well. According to analysts at MS Advisory writing in July 2026, China’s social credit system in its current form has no unified nationwide score for individuals, with implementation more focused on corporate compliance than personal scoring. The TBI proposal, by contrast, explicitly targets individual taxpayer ranking.

The structural comparison that critics find most relevant is not one of intent but of architecture: both systems involve the aggregation of financial and identity data across government databases, processed algorithmically to produce behavioral risk assessments tied to a unique identifier. The stated purposes differ; the underlying data infrastructure does not.

What Governments Are Being Asked to Build

The TBI report is addressed to governments — particularly those in the developing world where, as the report acknowledges, existing datasets “are incomplete and fragmented.” The argument is that even imperfect data, when cross-referenced through AI, “can still reveal” meaningful patterns of fiscal risk.

What governments are being asked to construct, in practical terms, is an integrated financial surveillance infrastructure: digital ID systems that serve as universal identifiers, linked to tax, payroll, procurement, and customs databases, with an AI layer on top that continuously scores each registered individual for their likelihood of non-compliance.

TBI’s framing emphasizes fraud reduction, tax compliance, and fiscal efficiency. These are legitimate policy concerns. But the infrastructure required to achieve those goals — once built — is not narrowly limited to those uses. A system capable of ranking taxpayers by fiscal risk using cross-referenced government databases and a universal digital ID is, by definition, a system capable of considerably more.

Transparency and the Question of Consent

TBI’s digital ID explainer asserts that properly implemented systems use “privacy-preserving safeguards that minimise data collection, restrict data retention, and give you full control over which personal information is shared, and with whom.” It also states that digital ID “is not designed to enable centralised tracking of citizens’ data and activity.”

The July 2026 fiscal report, however, describes a system in which government databases containing payroll, customs, procurement, tax return, and digital ID data are actively cross-referenced by AI to assess each individual’s financial behavior — before they have even filed a declaration. These two positions exist in some tension with one another, and that tension deserves public deliberation rather than resolution by policy institutes alone.

The Broader Policy Conversation

The Tony Blair Institute is an influential policy body with reach across multiple governments. Its recommendations carry weight. The July 2026 report is not a fringe document — it is a detailed, technically sophisticated proposal addressed to finance ministries and development institutions worldwide.

The question it raises is not whether governments should reduce tax evasion or cut fiscal waste. It is whether the infrastructure required to achieve those goals through AI-driven taxpayer ranking and universal digital ID represents an acceptable trade-off — and who gets to make that determination. As of the report’s publication, that conversation is largely happening inside policy institutions rather than in public legislatures.

The architecture TBI describes, once deployed, would be extraordinarily difficult to dismantle. That alone makes the proposal worthy of scrutiny that extends well beyond the fiscal efficiency arguments its authors advance.

This article draws on reporting from Activist Post / The Sociable, the Tony Blair Institute for Global Change, TBI’s Digital ID explainer, Wikipedia’s overview of China’s Social Credit System, and MS Advisory’s 2026 analysis of China’s Social Credit System.