In a recent discussion between Robert Goulder of Tax Notes and Lyla Latif of Warwick Law School, the importance of considering artificial intelligence (AI) in global tax reform was highlighted. The United Nations is working towards creating a more inclusive, equitable, and effective global tax system, and the integration of AI into this agenda is seen as crucial. Latif, a tax specialist, emphasized the significant impact that AI is having on reshaping business models across the global economy, presenting challenges for international tax policy.
One of the key challenges discussed by Latif is the reliance of AI-driven businesses on data and the complex nature of value creation in multiple jurisdictions. This makes it difficult to determine the value created by data and allocate taxing rights effectively. The development and deployment of AI systems involve various stages across different countries, leading to a fragmented value chain that is not adequately addressed by current international tax norms.
Latif also highlighted the reliance of AI-driven businesses on intangible assets such as algorithms, software, and intellectual property, which are key drivers of value creation in the AI economy. However, existing tax norms primarily designed for tangible assets may struggle to address the valuation and taxation of intangibles, leading to profit misallocation and erosion of tax bases.
The concept of “algorithmic colonization” was introduced by Latif to describe the potential consequences of uneven development and deployment of AI across the globe, particularly impacting the Global South. Large tech companies from developed countries concentrating economic power through AI technologies can lead to economic disparities and exploitation of data from developing nations.
To address these challenges, Latif proposed the idea of an AI-enabled Permanent Establishment (PE) as a new standard and norm for taxing AI activities. This approach would establish a threshold for taxable presence in a country based on AI activities such as data collection, algorithm development, and value creation, even without a physical office or employees present. This could provide a fairer distribution of profits and benefits of AI activities, particularly benefiting countries in the Global South.
The implementation of the AI-enabled PE concept could be achieved through bilateral or regional agreements or changes to domestic tax laws, offering a complementary approach to existing initiatives like the OECD’s Pillar 1 proposal. While Pillar 1 aims to address tax challenges related to digitalization, it may not fully capture the specific characteristics of the AI economy, making the AI-enabled PE a modern solution for modern times. The discussion concluded with a call for the UN to prioritize AI and data in the global tax agenda to ensure a more equitable and effective tax system tailored to the challenges of AI-driven businesses.
