Artificial Intelligence (AI) Washing in Taxation: Ethics and Transparency

Artificial intelligence (AI) is a broad term that refers to techniques making machines "Intelligent". Pascal A. Bizarro and Margaret Dorian (2017) pointed out that AI was introduced in 1948 when William Gray Walter created two small robots, named "Elmer" and "Elsie", that were able to recognize and respond to stimuli, encountering obstacles¹. Two years later, Alan Turing (1950) proposed that a machine could transmit information, communicate, and possess thinking capabilities indistinguishable from those of humans². In 1956, the Dartmouth workshop proposed the term "artificial intelligence", marking the birth of AI as a discipline. Since then, the AI phenomenon has received considerable attention in various fields. Over recent years, there has been a dramatic increase in the adoption of AI within the tax sector. This surge is attributed to improvements in algorithmic capabilities, greater computer power, and access to richer datasets. These advancements have enabled tax professionals to leverage AI for more sophisticated tax analytics and decision-making.

The field of deep learning gained popularity in the 2000s under the direction of researchers like Geoffrey Hinton, Yann LeCun, and Yoshua Bengi, which led to important advances in areas like image identification and natural language processing. These days, artificial intelligence is included in commonplace devices like GPT (Generative Pre-trained Transformer), autonomous automobiles, and virtual personal assistants for customer support³.

The term "AI washing" is the practice of exaggerating or falsifying the application of artificial intelligence (AI) in goods, services, or solutions to make it seem more sophisticated. This phrase, which comes from the word "greenwashing," is widely used in several areas, including taxes¹. The first use of AI in taxation was in Australia, where the Australian Taxation Office (ATO) started experimenting with AI technologies in the early 1990s. The ATO used AI to automate and streamline tax processes, particularly focusing on identifying fraud, tax evasion, and errors in tax returns. These early systems employed rule-based AI and later evolved to use machine learning techniques. Other countries like the United States and the United Kingdom followed suit, with agencies like the Internal Revenue Service (IRS) and HM Revenue & Customs (HMRC) integrating AI for fraud detection, auditing, and predictive analytics in tax compliance. As per the latest report of Thomson Reuters Institute, in 2025, around 21% of tax and accounting firms are either using or planning to adopt AI solutions, particularly for automating routine tasks like data gathering, compliance, and document processing.

The term "AI washing" is the practice of exaggerating or falsifying the application of artificial intelligence (AI) in goods, services, or solutions to make it seem more sophisticated.

Tax avoidance, the complex nature of tax regulations, and the high expenses related to their management and compliance make it difficult for fair taxation of different firms and lower-income groups. Automation and AI are being utilized more and more to address these problems. Tax-related software or services make claims about using advanced AI in the process of tax preparation, escalation, and fraud detection, but depend on simple algorithms and analytical methods.

This fraud may have serious consequences. Businesses and taxpayers may place a lot of confidence in the effectiveness of AI-driven tax solutions, which might lead to poor decision-making and excessive dependence on faulty systems, increasing exposure to risks³. Moreover, companies that make false promises about AI may face scrutiny from regulators and legal penalties if they are found guilty of misleading customers and tax authorities. Companies and the Board of Directors need to be transparent about the true abilities of their AI systems for establishing and retaining trust. Regulatory frameworks must be established to assess and check AI claims as AI becomes more integrated into tax systems.

Problem Statement: AI washing in taxation diminishes shareholders' trust, leads to potential regulatory breaches, and raises major ethical concerns.

Objectives

• To comprehend the concept of AI washing and its effect on taxation.
• To explore the ethical uses of AI in taxation.
• To review case studies of AI washing in taxation.

Literature Review

In the more recent decade of the 2010s, the practice referred to as "AI washing" became a major problem. The term was addressed in Forrester Research's report by Elizabeth Cullen in 2017. It refers to companies mislabelling or overstating their use of AI to capitalize and grow their companies. This term was introduced in a report to highlight the issue where businesses use AI labels to get market attention even when they don't involve AI technologies in their products or services. Several industries are using basic algorithms or basic automation, including technology, banking, and healthcare, and have begun advertising their goods as AI-driven. This resulted in the rise of AI washing. Some support platforms marketed their chatbots as highly advanced artificial intelligence (AI) systems, despite the fact that they relied mostly on pre-designed responses and rule-driven interactions.

Paschen et al. (2020) point out that AI washing in taxes can result in a dependency on inefficient systems, which may hinder decision-making and expose taxpayers and companies to more risk. False claims about AI can potentially damage public conviction in tax systems and AI technology, which can impact regulatory positions and reduce faith in AI-driven solutions⁴.

AI has black-box technology that makes it simpler for businesses to get involved in AI washing in taxes by hiding the real functioning of their systems. They can exaggerate AI capabilities without providing transparency, which makes it challenging for consumers or authorities to confirm whether AI is being applied to their services⁵. Strict rules and regulations are required for verifying AI claims and opposing AI washing. According to Binns, R. (2018), standards or certifications should be established to confirm AI technology and ensure that AI marketing is transparent. Some methods are critical in combating AI washing. Hassija explained explainable AI (XAI) in his paper, which helps in the decision-making process and provides an insight into how decisions related to certain tax filings were made by the system. By enhancing transparency, the key variables or data points in the model are revealed⁶.

Auditable logs can track every decision made by the AI system, which helps taxpayers or tax authorities to review the process and understand the reason why AI drew the specific conclusion. Akpan stated in his paper that Human-in-the-loop (HITL) is another strategy that is useful for involving human auditors. When black-box AI performs the bulk of the work, it can review and validate decisions for auditors to ensure fairness and transparency that impact taxpayers⁷. For building trust, open-source algorithms are used in which the functionality of the complex AI model is explained to make the system transparent.

Methodology

This study examines how AI washing is used in tax systems and the factors that impact online tax systems by analysing secondary data sourced from Scopus, Elsevier, Emerald papers, and peer-reviewed journals. The data collection process specifically targets recent articles published within the past decade, selected based on criteria such as relevance, peer-review evaluation, and the journal's impact factor.

Findings

The range of AI Washing in Taxation

Exaggerating the complex nature of AI solutions to draw in investment, boost one's reputation, or defend policy choices is a common phenomenon. Surveys and studies on AI adoption often reveal discrepancies between reported AI capabilities and actual implementations. For example, The Financial Express claims that Venture capitalists are increasingly concerned about AI washing, where up to 70% Startup companies falsely claim AI capabilities to secure funding, as depicted in Figure 1. This problem is also evident in the taxable domain, where private companies and government tax authorities exploit the excitement around artificial intelligence to promote efficiency and innovation.

Prevalence of AI Washing Practices

The ability to detect tax fraud and improve cooperation has been a major advantage for tax authorities worldwide, who have embraced AI technology quickly. However, these claims are frequently not realised in practice in the real world. While AI techniques have been integrated into tax systems, research by the International Monetary Fund (IMF) suggests that their influence has been somewhat limited in comparison to the expectations set by public statements. Similarly, companies have been found to overstate the contribution of AI in their tax management procedures, captivating investors with creative concepts that are not adequately supported in their daily activities⁸.

Common Methods and Mechanisms Used in AI Washing

In the context of taxation, the use of vague language and exaggerated success claims are two key indicators of AI washing. Often, companies refer to their systems as "AI-powered" without clarifying what proportion of the procedure has been automated or how much still relies on human monitoring. Even though the actual system primarily depends on traditional rule-based techniques improved with some machine learning algorithms, term such as "AI-driven fraud detection" is used. Highlighting specific success stories while minimizing deeper systemic flaws is another popular strategy. The true potential and readiness of AI systems for taxes are misrepresented to participants through these operations.

Impacts on Tax Authorities and Public Trust

The complexity of AI technology makes it challenging for tax authorities to identify and control AI washing. As AI is developing so quickly, it is challenging for regulators and policymakers to keep up with the latest technologies and correctly determine the genuine capabilities of AI systems. There are further complications that arise due to the absence of universal norms and accurate definitions of what AI is doing, which may result from regulatory agencies' frequent lack of the expertise needed to examine AI claims carefully.

Public trust, including shareholders in the organization, may decline if they discover that the AI capabilities have been overstated, affecting both investors and consumers.

Companies that employ AI washing risk serious legal problems and harm to their reputation. Public trust, including shareholders in the organization, may decline if they discover that the AI capabilities have been overstated, affecting both investors and consumers. This breakdown of confidence can lead to loss of economic potential, market value, and possible legal implications, ultimately decreasing public trust in AI technology and impeding innovation and wider adoption. Overstated AI claims generate mistrust and inflated expectations, which hinder the development of innovative AI applications.

Case Studies

Unethical Implications of AI in Taxation

Misinterpretation of Data in Tax Filing Services

A business offering tax filing services faced allegations of misleading its clients and taking advantage of them by charging for services even when they were eligible to get them for free using the IRS (Internal Revenue Service) Free File program. The allegations have been made against the company, claiming that they purposefully diverted users from the free alternatives to the premium products to gain from AI-enhanced marketing strategies rather than AI-driven tax preparation advantages.

The IRS Free File program makes partnerships with entirely-profit tax software providers like Intuit, allowing qualified taxpayers to electronically prepare and file their taxes for free⁹.

Misleading Information by AI

Some of the tax preparation firms lied about their capabilities of AI-powered tax preparation software. There have been accusations of "AI washing" as a result of customers' and analysts' anticipations that the real benefits of AI could not match the marketing claims. Due to these claims, consumer advocacy groups have launched legal challenges to ensure accuracy and transparency in the marketing of AI technology used in tax preparation services¹⁰.

Exaggeration of AI capabilities in financial services

Some of the renowned credit score monitoring and financial services have been accused of AI washing about its tax preparation services. The company has expanded its services by including AI-driven tax preparation tools, which have been promoted as convenient and reliable options for users by making the tax filing process easier. It has been accused that it is a part of a marketing strategy to attract users and gather data. Its use of fraudulent and unethical marketing strategies has led to lawsuits against it, as well as demands for an investigation and legal action against the company's activities⁷.

Ethical Implications of AI Washing in Taxation

Privacy and Data Security

AI tax systems follow strict data protection guidelines to prevent breaches and unauthorized access, as they require financial and personal data of the users to function, but this data should be recorded ethically. To preserve taxpayer privacy, it is essential to make sure about data privacy.

Bias and impartiality

An AI system may make judgments that unfairly affect particular taxpayer groups if it is educated on past data that contains biases. Maintaining justice in tax administration requires making sure AI technologies are developed and evaluated to reduce bias. In 2017, the Income Tax Department of India used AI and data analytics in its online taxation system for tax investigation, reducing human intervention and subjective bias. AI chatbots used by income tax departments for solving taxation queries, called tax bots, highly influence taxpayers to make unbiased and impartial decisions while paying taxes.

Transparency and Accountability

The "black-box" nature of AI systems show challenges to transparency and accountability. It is often difficult to understand how AI makes decisions or recommendations. To address this, tax authorities need to provide clear interpretations of how AI systems function and the criteria used in decision-making processes. Explainable AI, auditable logs, Human-in-the-loop (HITL), and Open-Source Algorithms are some strategies that can be used to maintain transparency. Transparency helps build trust and allows stakeholders to hold institutions accountable for errors or unfair practices.

Ethical Use of AI for Compliance and Enforcement

As AI can improve revenue collection and regulation, there are moral concerns around the application of these technologies. For example, employing AI to actively investigate tax evasion may result in taxpayers being treated unfairly or under excessive scrutiny. It's critical to strike a balance between the advantages of AI in enhancing compliance and the need to protect taxpayer rights and avoid excessive enforcement.

Impact on Employment

AI adoption in tax administration has the potential to significantly alter employment patterns. Although AI can save administrative costs and simplify processes, it can also result in the loss of tax professionals' jobs. Providing support and opportunities for skill upgrading to impacted employees is an ethical consideration that should be prioritized to ensure a fair transition and minimize adverse effects on the workforce.

AI tax systems follow strict data protection guidelines to prevent breaches and unauthorized access, as they require financial and personal data of the users to function, but this data should be recorded ethically.

Regulatory Measures in India

In India, the Digital Personal Data Protection Act (DPDPA) is an act passed in August 2023 that regulates how personal data is collected, processed, and used in a fair, transparent, and accountable manner. It sets guidelines for AI systems handling personal and financial data, including that data which is used in taxation¹⁰. India's National Strategy for Artificial Intelligence, released by NITI Aayog, outlines the government's approach to AI, including the promotion of responsible and ethical AI practices by large language models (LLMs). While it is not specific to taxation, it sets a framework for AI development and implementation. In India, the CBDT is in charge of tax administration and regulation. While there are currently no formal standards and regulations on AI washing, AI systems, and tools used in tax operations would be bound to the CBDT's standards on tax preparation and reporting.

Conclusion

AI washing in taxation undermines trust and efficacy in tax systems, leading to poor decision-making and potential legal issues. To combat this, clear regulations, transparency, and ethical AI practices are essential. Ensuring accurate representation of AI capabilities and educating consumers can help maintain trust and effectiveness in AI-driven tax solutions.