The Untapped Potential of AI for African Revenue Authorities

What is artificial intelligence, and can machines think? In 1950, Alan Turing created a test to evaluate the intelligence of a machine. This test, that came to be known as the “Turing Test,” consisted of an evaluator, a human respondent, and a machine. The evaluator had to ask questions and classify a machine as intelligent when the evaluator could not differentiate the answer of the machine from that of the human.

The Organization for Economic Co-operation and Development defines artificial intelligence as a machine-based system that can, “for a given set of human-defined objectives, make predictions, recommendations or decisions influencing real or virtual environments.” Yet what is the current status of adoption of AI technology?

Digital trends such as AI are transforming economies, governments and societies; such a technological change is altering the way people interact with their governments and requires a shift in the design and delivery of public services. Governments will need to use digital technologies to deliver timely proactive and inclusive public services. The innovative and collaborative approaches brought by digitalization are creating greater trust in public institutions.

However, the AI disruption brought about by digitalization is also impacting the job market. The OECD considers that 9% of jobs in member countries are going to be automated, while an additional 25% will be deeply altered. In that context, Africa is more seriously impacted by the adoption of AI solutions and automation: 85% of jobs in Ethiopia are at risk of being replaced by automation, while the figure is 67% in South Africa.

In the context of tax revenue mobilization and collection, we will explain below how AI is an opportunity for African tax authorities, but why several key challenges need to be addressed before any further development occurs.

Opportunities and Potential of AI in an African Context

Recent Acceleration in AI

Since the Turing test, and since John McCarthy defined AI and is considered its father, AI has progressed considerably. The AI environment evolved with the Internet of Things and Industry 4.0, where huge amounts of data have become available from electronic devices, sensors, appliances, machines, and vehicles.

The availability of big data has fueled the development of machine learning and deep learning in the 1990s and into the 21st century. Thus, there is no need for a human to feed machines with rules, as machines are learning by themselves from the huge amounts of data available, and creating new perspectives.

In the area of public sector services, AI can help governments design better policies, make better decisions, optimize communication with all stakeholders, increase the speed of delivery of services to citizens, and help shift the work of civil servants from routine work to high-value tasks.

In Africa, AI is already being used in conflict-prone areas on the battlefield to fight insurgencies, rebels, and terrorists. But what are the other fields where AI is used?

A Tool to Leverage Development Projects

On Oct. 26, 2019, the African ministers in charge of information and communication technologies adopted the African Digital Transformation Strategy. The DTS for Africa aims at driving digital transformation, and cross cutting themes to support the digital ecosystem.

One of the recommendations of the DTS was the establishment of a working group on AI to study the creation of a common African stance, with capacity building and targeted projects.

The first session of the Africa Union working group on artificial intelligence took place in Cairo in December 2019, where priorities on AI for Africa were determined and where a discussion took place about using AI to overcome development obstacles and challenges.

This first session was followed by a second session in February 2021, where the working group stressed the importance of exchanging experiences and cooperating to bridge the digital divide between developed and African countries. The working group also stressed the importance of developing a joint capacity building framework in Africa to foster digital and AI development.

Importance of Digital Foundations in Africa

Yet the application of AI is not a reality in most African countries, except for Kenya, South Africa, Nigeria, Ghana, and Ethiopia. The main success factors necessary for AI to develop in Africa are generally not present, such as reforms covering data collection and data privacy, infrastructure, education, and governance.

The above prerequisite reforms are essential for AI to promote “sustainable development and inclusive growth.” They are the reasons why African countries are struggling to meet the requirements for AI to perform. These requirements encompass digital foundations, including the availability of large volumes of data—big data—and specific competencies in human adopters, which most African countries currently struggle to meet.

Two Proposals for African Tax Authorities Using AI

Addressing the Legal Challenges of AI

As a new tool capable of changing the way that tax law is practiced, revenue authorities will be wise to investigate the risks induced by such an integration of AI into tax law, first of all legal risks.

AI in fact poses a threat to the sustainability of tax systems, despite the multiple technology opportunities it offers. AI integrated in tax systems involves processing personal data of taxpayers, disclosed by taxpayers or collected by tax authorities from other sources, and calls for consideration of the General Data Protection Regulation. The GDPR will also apply when revenue authorities apply AI in profiling and automated decision-making.

So far, soft and hard law instruments on AI may not sufficiently address the particular information needs of the tax domain. Adequate protection of taxpayers’ rights requires the use of eXplainable AI that can render the functioning and decisions of tax AI systems understandable for taxpayers, administrative appeal bodies, and courts. Tax authorities are constitutionally responsible for miscalculations of taxes or misidentification of tax risks resulting from their use. Such mistakes may follow from coding errors (bugs).

A New Source of Tax Revenue—AI and Robots

It is a recent proposal that AI should pay taxes, considering that tax systems tend to incentivize automation while tax revenues may still derive from labor income. Yet AI replaces human workers and is likely in the long term to generate a loss of payroll tax for governments. Tax systems must then be progressively redesigned to tax capital rather than labor.

To a lesser extent, the adoption of an “automation tax” could help to slow the adoption of automation technologies, giving workers and social support systems time to adapt. An automation tax could rely on changes to existing rules of depreciation or capital allowances, that will reduce the tax incentives attached to AI development and implementation.

To a larger extent, it is possible to tax robots (i) as a form of consumption of commodity or service, (ii) for the income generated as a factor of production, (iii) to allow depreciation from the income tax base for depletion of value, or (iv) as a combination of all three, which is more likely to occur. From a policy perspective, the possibilities of taxing robots seem numerous, but the design of robot taxation may rest on how it is perceived—whether as labor or capital—though both forms may be contemplated.

One should keep in mind that AI is a natural tax avoider. If there is too close of a connection to a taxing jurisdiction, it is simple enough to be modified to remain non-taxable in that jurisdiction. AI taxation will inevitably question in the future the definition of “permanent establishment.” To be continued …

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

Author Information

Anthony Assassa is an associate member of the Chartered Institute for Securities & Investment with more than 12 years’ Africa and Asia background (Cameroon, Comoros, Congo DRC, Laos) in reforms towards mobilization and collection of tax and customs revenues. He attended the international specialization cycle in tax and customs administration of the École Nationale d’Administration, Paris.

Elie Sawaya led several governmental reforms in Asia and Africa and is a digital, private and public sector expert working for GIZ with more than 20 years of practice in public private dialogue, e-government, private sector development, port community systems, supply chain, international trade facilitation, electronic tax systems, risk management, customs and cross border trade.

The authors may be contacted at: anthony.assassa@gmail.com; elie.y.sawaya@gmail.com