Companies and the public sector are aiming for efficiency and savings through the use of AI. This applies also for tax matters. While this development is welcome, it must be ensured that data is not used carelessly; taxpayers’ rights, data protection, and transparency of the administrations’ actions must always be safeguarded.
To develop AI and new operating models, AI investments should also be supported with tax incentives. Currently, AI, software, and services are not sufficiently recognized as R&D activities, so investments risk being excluded from research and development tax deductions.
Goals: efficiency, reduced administrative burden, and easy access to high-quality public sector services
The use of AI in taxation has become a hot topic in recent years, both in the business and the public sector. Tax reporting has increased exponentially over the past ten years, for example due to minimum tax legislation. With the EU’s ViDA directive (VAT in the Digital Age) coming into force in 2030, VAT reporting will be based on vast amounts EU standard e-invoicing data. Mapping and analyzing this extensive amount of data will not be possible without AI-based tools.
I had the privilege of talking in the 2024 IFA Congress WIN-panel, with the topic “Tax and Technology: Innovative Use Cases from Around the Globe.” representing the continent of Europe. More on this topic can be found in Verotus-lehti issue 1/2025, which lists various ways tax authorities all around the world use AI. This year’s IFA Congress will likely also touch on use of AI in the lectures on current EU and OECD tax projects and on reducing administrative burden.
For companies, it is important that tax authorities worldwide use AI not only to improve service levels—such as virtual assistants and automated decision-making processes—but also to speed up advance rulings and tax audits. Utilizing AI in taxation can significantly reduce companies’ administrative burden and bring savings, as human errors decrease and the accuracy of tax processes improves.
Data collection and use must not be reckless
The EU’s General Data Protection Regulation (GDPR) requires that only strictly necessary, precise, and proportionate information for taxation purposes can be requested from taxpayers. Data cannot be requested indefinitely without knowing its intended use, or even more alarming, if data is requested without a purpose or need to use it. “More data does not mean more taxes.” Taxpayer rights, data protection, fairness, transparency, and responsibility are crucial.
Pending legislation and recent news about unnecessarily broad data requests by the Tax Administration are causing concern among taxpayers.
- Pending legislation on comparative data tax audits would give the Tax Administration authority to collect mass data. Based on the law, a company could face a situation where it must choose between violating tax law or breaching constitutionally protected data and GDPR.
- In April, the Helsinki District Court sentenced two tax auditors to fines for improperly requesting and using telephone connection and location data from a telecom operator during a tax audit.
- Tax auditors have demanded confidential and sensitive patient records and photos from plastic surgeons (plastic surgery procedures are VAT-exempt if medically necessary; otherwise, the VAT rate is 25.5%).
The OECD’s Tax Administration 3.0 project promotes reliable use of AI by tax authorities. Cooperation between countries, tax administrations, companies, and AI tool (software) providers play a key role in ensuring the tax systems are compatible and national deviations in implementing legislation are avoided.
Development is hindered because R&D tax incentives don’t recognize AI investments
Companies are actively developing AI-based solutions to enhance their operations and remain competitive in a rapidly changing world. AI is used for process automation, improving customer service, optimizing financial management, and risk management. By developing their own AI solutions, companies can analyze large amounts of data more efficiently, identify new business opportunities, and respond to market changes faster.
To promote this, Finland offers a tax deduction to encourage companies’ own R&D activities. However, it is concerning that the OECD’s Frascati Manual—which is used by the Finnish Tax Administration to define research and development—does not mention AI at all. This means AI-related investments may be excluded from R&D tax incentives, even if they are innovative and development-oriented. The Frascati Manual dates from 2015 and the supplementary Oslo Manual from 2018.
Finland should expand the definition of R&D to better cover AI, software, and service investments, and actively participate in the OECD’s ongoing update of the Frascati Manual. Taxation and public support should also follow OECD’s recommendation to include AI in the definition of R&D. Many countries support companies investing in AI with various tax benefits, and/or AI is included in the definition of R&D, making it eligible for R&D tax incentives.
AI use and development cannot overlook need for talent
In addition to interpretation challenges with the Tax Administration’s R&D tax incentive, companies’ and the public sector’s AI investments are also slowed by many practical obstacles. Data availability, quality, and correct format are often lacking, and resources—funding, skilled employees, and time—are often limited. However, one challenge stands out: the availability of skilled professionals.
Who is responsible for training new AI experts? Should individuals acquire AI skills alongside their own expertise, or should training be included in educational degrees? Is training the responsibility of the employer—public or private sector? Basic skills are no longer enough when AI experts are needed for decision-making, system building and use, auditing and correction, and training new experts.
Utilizing AI in taxation and society more broadly thus requires not only technological solutions but also strong investment in skills and education. Only in this way can we ensure that AI’s full potential is realized—responsibly and efficiently. Attracting and rewarding talent can also be influenced by taxation, for example by lowering income tax and top marginal tax rates.
About the writer
