An AI-Powered Automated Tax Code Must Uphold Core Principles

With smarter algorithms, tons of data, and powerful computers, digital tax systems are accelerating.

According to the OECD’s 2024 survey, over 70% of tax administrations are using AI to work more efficiently, catch fraud, and improve services for taxpayers. More than 80% are building application programming interfaces—sets of rules and protocols that allow different software systems to communicate and interact—to automate and streamline how tax data is shared.

Estonia uses risk detection models and a tax behavior rating that compares taxpayer data with others and public records. Chile uses tech solutions to make tax compliance easier, having used AI to answer tax questions and recover millions of dollars. And the IRS aims to use AI to improve audits and data streams to identify returns for audit.

The next technological step could be huge: fully automated tax systems. The challenge lies in ensuring that any automated system respects democratic principles and constitutional foundations.

Automated Tax Code

Imagine a world where the tax code isn’t printed in books, debated in parliaments, or enforced by auditors. Instead, the tax “code” lives in the cloud—self-executed and tweaking itself constantly to meet economic goals. Every transaction pays its fair share automatically, with no loopholes in sight.

While this vision promises unprecedented efficiency, it also introduces significant threats. The question is how it should be designed and governed, not whether it’s possible.

Core values of democracy are at stake: who protects freedom in a system administered by a code, who ensures fairness when codes are biased, and how transparency survives in the opacity of algorithms.

A Constitutional Question

In Gregory v. Helvering, the US Supreme Court recognized a foundational principle: Taxpayers may arrange their affairs to minimize tax liability, provided their actions have economic substance beyond mere avoidance.

This ruling enshrined a vital space for the economic right to organize affairs within the tax system—a space where individuals and businesses can exercise judgment, flexibility, and discretion to navigate complex financial realities. Several other US Supreme Court rulings reinforced and expanded this principle.

In a fully automated tax system, that discretion would be compromised. Where taxpayers once relied on professional judgment to interpret gray areas such as transaction classification, tax credit eligibility, or residency, algorithms would execute rules mechanically. Complex realities could be flattened into binary outcomes, with no room for debate over intent or economic substance.

Freedom at Stake

As automation increases, the line between administering policy and defining it blurs. Algorithms embedded in the tax code might systematically misclassify certain business models, wrongly reject legitimate transactions, or deny valid deductions. In analog systems, such errors can be challenged through audit, appeal, or litigation. In real-time, code-governed systems, the process is less transparent—and far harder to contest.

Tax systems have always done more than collect revenue—they are tools of policy. They nudge us to save, reduce emissions, and shift consumption.

In the analog world, this influence was indirect and contestable. In the digital economy, it becomes direct, immediate, and embedded in infrastructure. Just as a self-driving car can’t break the speed limit, or a smart fridge won’t open if it restricts access based on your diet settings, a purely digital system would lock taxpayers into a system that controls behavior rather than interprets economic reality.

Accountable Tax Code

None of this should be mistaken as a rejection of digital technologies applied to taxation. The automation of tax processes offers extraordinary potential for fairness, efficiency, and transparency.

That said, building and enforcing a framework for accountable automation is no simple task. We face technical, legal, ethical, and political challenges. To harness the benefits of automation while preserving legal and democratic values, any automated tax code should encapsulate several core principles.

Algorithmic transparency. Taxpayers should understand how automated decisions are made. Each action should include a plain-language explanation, and tax authorities should publish real-time dashboards to show how decisions are applied.

Independent oversight and preclearance. Algorithms with significant tax implications should undergo review by an independent authority before deployment. They should also be subject to regular, third-party audits to evaluate fairness, legality, and accuracy. Whistleblower protections should cover those who report flaws or misuse.

Appeal mechanisms. Taxpayers must have a clear path to contest automated decisions before a tribunal or court. That process depends on reliable decision logs and audit trails documenting the rules and data used.

Accountability. Responsibility for the fairness and legality of an automated tax code must lie with those who design, deploy, and operate them. This includes codes of conduct for engineers and public sector officials, as well as strict liability for harm caused by algorithmic decisions.

Democratic participation: Taxpayers and civil society should be given opportunities for public comment on proposed algorithmic rules, access to simplified models that explain how systems work, and clear channels to report errors or unfairness.

Cybersecurity and privacy. Privacy is fundamental, not optional. As tax systems grow more automated and data-heavy, protecting sensitive information from abuse and cyber threats is essential to safeguard rights and sustain public trust.

To achieve this, systems should use end-to-end encryption, adopt zero-trust architectures—which eliminate implicit trust and require constant authentication—to reduce breach risks, and undergo regular independent audits to identify and fix vulnerabilities.

These accountability principles can’t be fully realized within any one jurisdiction alone. In our interconnected world, where tax data and financial transactions flow seamlessly across borders, divergent standards create loopholes, enable abuses, and open doors to regulatory arbitrage. True accountability requires international cooperation.

Looking Ahead

The future of taxation isn’t just about better tools; it’s about who controls them and how decisions are made.

Tim Berners-Lee, the inventor of the World Wide Web, said that those who build digital systems aren’t just engineers; they are architects of new worlds. As the global tax community moves toward more intelligent and responsive systems, we must ensure that efficiency doesn’t come at the cost of freedom, fairness, transparency, or democratic legitimacy.

This article does not necessarily reflect the opinion of Bloomberg Industry Group, Inc., the publisher of Bloomberg Law, Bloomberg Tax, and Bloomberg Government, or its owners.

Author Information

Sofía Larrea, an international tax attorney based in Ecuador, specializes in US and cross-border taxation and is admitted in Ecuador, California, and the US Tax Court.

Write for Us: Author Guidelines