How AI is changing the work of tax authorities and impacting taxpayers?

Automation of tax processes using artificial intelligence (AI) is becoming increasingly common among businesses as well as tax authorities.

Tax authorities are benefiting from the technological revolution

From the perspective of tax authorities, AI first and foremost allows for more efficient analysis of the vast amounts of data received from taxpayers. This allows for faster detection of irregularities and potential tax fraud. Many countries are also testing solutions such as chatbots and virtual assistants to help taxpayers solve simpler problems and answer questions on their tax returns.

According to an OECD report, more than 80% of tax administrations are already using or are in the process of implementing in the future technologies based on artificial intelligence and machine learning to use data in a way that reduces the need for human intervention. An example of the use of analytical tools can be seen in Poland’s STIR system (Teleinformatics System of the Clearing House), which analyses banking operations in real time, identifying high-risk transactions and informing the tax authorities about them.

Productivity gains through AI can contribute to the reallocation of human resources to more complex tasks that require creative thinking and interpretation of regulations. Tax administration staff can focus on analysing complex and ambiguous cases while routine tasks are taken over by AI systems.

Challenges and risks associated with the use of AI by tax authorities

No process of implementing innovative solutions is without its difficulties. One of the key challenges related to the use of AI by tax authorities is the so-called black box effect, i.e. maintaining transparency in AI-based decision-making processes. This is because taxpayers should be able to understand how and on what basis decisions are made regarding their tax liabilities. However, it is often difficult to explain how AI arrived at a particular decision or conclusion. This raises concerns about the possibility of taxpayers challenging such decisions and the potential infringement of their right to a fair resolution of their case.

In this context, it is worth noting the solution introduced in France, among others, where citizens are guaranteed access to information on the use of algorithms in individual administrative decisions, including how they work and the weights assigned to specific data used in the process.

Another important aspect is to ensure non-discrimination and fairness in AI solutions used by tax authorities. Algorithms must be designed and trained to avoid bias and unjustified favouritism towards certain groups of taxpayers. This requires careful selection of training data and regular audits of AI systems for potential discrimination.

Privacy and data security are another significant challenge. Tax authorities need to ensure adequate safeguards for the vast amounts of sensitive data processed by AI systems. Advanced encryption mechanisms, access controls and monitoring of potential security breaches need to be implemented.

The issue of accountability for decisions made by AI systems is also worth noting. In the event of errors or irregularities resulting from the operation of the algorithms, there must be a clearly defined redress path for taxpayers and corrective mechanisms on the part of the tax authorities.

In its recommendations on the use of AI in public administration, the OECD stressed the importance of an ethical approach to the implementation of these technologies. To ensure this, it recommended, among other things, retaining human oversight of key decisions made by AI systems, especially in cases that could have a significant impact on taxpayers’ rights and obligations.

How can taxpayers prepare for automation in authorities?

Taxpayers should not be left behind in the automation race observed worldwide. Saving the time needed to handle and settle taxes can bring them even clearer benefits (including financial ones related to accounting and tax service costs) than tax authorities. A key element of this process is the digitalisation and automation of internal tax and invoicing processes.

However, when implementing new solutions, it is important not to forget the requirements that will be placed on taxpayers by planned invoicing revolutions such as the National e-Invoicing System (KSeF) in Poland or the proposal for near real-time reporting of intra-Community transactions put forward as part of the EU’s VAT in the Digital Age (ViDA) package. Taxpayers should therefore exercise caution not to invest resources in systems that may prove incompatible with the new regulations. It is also equally important to raise competence in new technologies and their impact on tax processes. This applies to both IT and finance or accounting staff, who use the new solutions in practice.

The implementation of AI technology by tax authorities leads to increased transparency. Taxpayers must expect that tax authorities will have a much broader arsenal of tools to analyse their financial data. This, in turn, will enable faster identification of potential irregularities or tax avoidance attempts.

New technologies will also affect the way taxpayers and tax authorities interact. We can expect more automated communication, which, on the one hand, may improve contact; on the other hand, it is not difficult to fear that interactions lacking empathy and contextual understanding may lead to frustration for taxpayers. In addition, an over-reliance on automated systems may lead to a dehumanisation of the relationship between taxpayer and tax administration, making it more difficult to deal with more complex issues that require human judgement and flexibility.

With the increasing use of AI in decision-making processes, taxpayers may face the challenge of challenging algorithm-based tax decisions. This may require the development of new legal and technological competencies within companies.

The issue of data security cannot be overlooked either. As more and more information is processed, the risk of potential leaks or unauthorised access to sensitive financial data increases. Therefore, investment in cyber security is becoming as important as the tax management tools themselves.

The future of tax in the age of AI

Looking ahead, we can expect to see the further development and integration of AI solutions into tax systems around the world, which will undoubtedly be followed by new challenges and issues.

The key issue in the future may be balancing efficiency and fairness in the tax system. While AI has the potential to significantly increase the efficiency of tax collection, it is important that this does not lead to an undue burden on taxpayers, a wider misunderstanding of the tax system than before, or simply a violation of taxpayers’ rights.

It will also be crucial to ensure that the benefits of using AI in tax are available not just to authorities or the largest businesses but to all taxpayers, regardless of their size or resources. This may require state support for small and medium-sized businesses in implementing new technological solutions.

[1] OECD (2023), Tax Administration 2023: Comparative Information on OECD and other Advanced and Emerging Economies, OECD Publishing, Paris, https://doi.org/10.1787/900b6382-en.

[2] https://guides.etalab.gouv.fr/algorithmes/