Artificial intelligence promises to revolutionize the global economy, but the distribution of its benefits and the taxation of this new wealth have become critical challenges for fiscal justice.
The exponential advancement of artificial intelligence (AI) projects a future where this technology will be intrinsic to production, labor, and the delivery of government services. However, the widespread use of AI does not, in itself, guarantee an equitable distribution of the wealth it generates. There is a real risk that the fruits of this new economy will become concentrated in the hands of a few conglomerates and a restricted number of nations, raising complex questions about fiscal justice and economic sovereignty.
The core concern lies in identifying where wealth generated by AI is actually created, who appropriates it, and how states can effectively capture its results. The simple use of AI tools by millions of users does not translate into direct economic participation in the profits. Recurring payments for licenses, computing power, and services tend to converge toward a few companies, often based in just a handful of countries.
Universal Access, Concentrated Wealth: The AI Dilemma
The democratization of access to AI tools can coexist with a significant concentration of their economic results. A Brazilian company, for example, may increase its productivity thanks to AI while simultaneously directing a considerable portion of its revenue to foreign suppliers. This dynamic is problematic when dependence deepens, the costs to replace these suppliers rise, and a small group begins to dictate the rules of access.
Distinguishing between access to technology, the capacity to develop it, and participation in its financial benefits is fundamental. The United Nations Conference on Trade and Development (UNCTAD) points out in its 2025 report that only about 100 companies, mostly from the United States and China, are responsible for 40% of global research and development investment. This data, while encompassing general business research, illustrates the asymmetry of resources shaping the technological race.
The decisive tax issue is to identify where wealth is formed, who appropriates it, and which states can effectively capture its results.
This concentration of wealth does not necessarily stem from asset concealment or artificial profit shifting. The ownership of AI models, control over technological infrastructure, financing capacity, and strategic positioning in distribution networks are factors that sustain a powerful economic concentration. Combating tax evasion remains a necessity, but fiscal justice requires going further by addressing the distribution of benefits from a globalized economy that is global in its use, but still restricted in its ownership.
Tax Reform and the Role of Cashback
Brazil’s tax reform, through Constitutional Amendment No. 132/2023 and subsequent supplementary laws, seeks to address some of these issues, especially regarding consumption. The introduction of mechanisms such as the cashback program, provided for in Supplementary Law No. 214/2025 and its amendments, emerges as an important redistributive tool. By returning a portion of taxes paid by low-income families, the cashback increases disposable income and mitigates the regressivity of consumption taxes.
Targeting this benefit toward vulnerable families enrolled in the CadÚnico program with limited per capita income distinguishes it from broader tax breaks. This mechanism aims to reduce the proportional tax burden for those who consume most of their income, acting as a concrete response to domestic inequality. However, the reach of the cashback is limited: it mitigates the regressivity of consumption but does not grant the population ownership of technological assets nor does it directly redistribute global AI profits.
Advances in International Taxation: Looking Beyond Borders
The Brazilian legal system has also sought to improve the taxation of income earned abroad by residents and companies. Law No. 14.754/2023, along with subsequent regulations, reorganized the taxation of financial investments, controlled entities, and trusts abroad, establishing clearer tax rates and reducing tax deferrals.
For companies, the introduction of a new transfer pricing regime under Law No. 14.596/2023 aims to correct distortions in intra-group transactions. The regulation of intangibles is particularly relevant for the taxation of payments and structures linked to technology. These advances seek to ensure that the Brazilian tax base is adequately considered, although they do not grant Brazil general jurisdiction over the profits of foreign conglomerates with global operations.
Global Cooperation and the Pursuit of Equitable Taxation
International concentration of income generated by AI becomes apparent when countries are major consumers of the technology but have little participation as owners of its assets or beneficiaries of its profits. Licensing revenue, taxable profit, and shareholder equity are not directly equivalent, and the jurisdiction where a conglomerate is headquartered may not coincide with the jurisdictions receiving its payments.
Initiatives like the OECD’s Pillar Two, which establishes a global minimum tax for large corporate groups, seek to combat profit shifting to low-tax jurisdictions. In Brazil, the Additional CSLL, aligned with this floor, is an example of this international coordination. However, the minimum tax does not constitute an automatic sharing of profits among all consumer countries.
The central question for technology-importing economies persists: how can they recognize their economic contribution to value generation without imposing arbitrary charges, double taxation, or international fragmentation? The answer lies in fiscal cooperation, consistent criteria for connection and attribution of results, and effective competition policy. Market concentration does not dissipate through tax collection alone.
Productivity Gains and the Need for Broad Participation
AI has the potential to boost productivity and create new opportunities, but ensuring that labor compensation keeps pace with this growth is not automatic. Fiscal policy must consider the possibility that technological expansion and the distribution of its gains may follow different trajectories. Fostering education, research, infrastructure, and competition is crucial for the country to participate actively in AI production rather than merely consuming external solutions.
Tax administration must also respect individual rights when using AI, as demonstrated by the Dutch experience with the SyRI system, which raised concerns about disproportionate surveillance. In Brazil, due process of law, the justification of decisions, and the General Data Protection Law (LGPD) require the possibility of contesting automated decisions and knowing the criteria used.
In short, Brazilian tax reform is advancing on several fronts, from consumption taxation to improvements in international taxation. However, the future challenge lies in ensuring that the wealth globally produced by AI finds legally sound ways to support opportunities and prosperity in the territories that actively participate in this technological transformation.