AI in Tax Law

Artificial Intelligence is prevalent in Tax analysis and planning but will it come to law?

Revenue Canada acknowledges it is using Artificial Intelligence: “Artificial intelligence (AI) technologies offer promise for improving how the Government of Canada serves Canadians. As we explore the use of AI in government programs and services, we are ensuring it is governed by clear values, ethics, and laws” (

What’s more, Revenue Canada has a mission statement and four guiding principles with respect to the use of AI: “The Government of Canada has released the Strategic Plan for Information Management (IM) and Information Technology (IT) 2017 to 2021, an update to the inaugural Government of Canada Information Technology Strategic Plan 2016‒2020, published in June 2016 ….

“It creates a framework and sets direction for the GC to become an open and service-oriented organization that provides programs and services to citizens and businesses in simple, modern and effective ways that are optimized for digital and available anytime, anywhere and from any device.

“Consistent with the GC’s first Strategic Plan, the following 4 strategic goals frame the direction for the GC: service, value, security and agility.

“Four strategic areas of action will achieve these goals over the next four years and beyond. Each area of focus — Service, Manage, Secure, and Community — details specific actions and activities that are underway or that represent new enterprise directions.
  •   Service focuses on building and evolving IM-IT foundational elements, including processes, practices and infrastructure, to enable implementation of current capabilities, technologies and solutions.
  •   Manage addresses how the management and governance of IM-IT across government ensure that IM-IT investments take advantage of economies of scale, demonstrate value and are sustainable.
  •   Security focuses on safeguarding sensitive government data and ensuring that Canadians who access online services can trust the government with their personal information.

  •   Community focuses on building a high-performing IM-IT workforce that has the skills and mindset needed to work effectively in an open digital environment and ensuring that public service employees have a modern workplace, professional development and the IM-IT tools they need to do their jobs.”

If Revenue Canada is using AI, what about taxpayers, especially corporate taxpayers? Let’s start with an academic premise: The abstract for academic Blazej Kuzniacki’s “The Marriage of Artificial Intelligence and Tax Law: Past, Present, and Future,” (available at SSRN: goes as follows:

“According to recent research’s prediction, global GDP could be up to 14% higher in 2030 as a result of various artificial intelligence (AI) applications, which is the equivalent of an additional $15.7 trillion. It makes AI oriented sectors the biggest commercial opportunity in the currently supersonic fast changing economy. This contribution, perhaps surprisingly, does not aim to propose how to tax profits generated by AI industries. The author rather takes an attempt to depict a potential of AI technologies to be applied to tax law. Let us see if AI can be happily married with tax law in order to get the best of both worlds.”

Arguably, Tax Law is one of the areas most susceptible to AI solutions. This may owe to the availability of AI solutions in Tax administration and management itself. As Deloitte LLP’s website says: “To manage the changing tax landscape, alongside the increased use of analytics, tax authorities and tax advisors are starting to explore the possibilities for deploying sophisticated data analytics and Artificial Intelligence (AI) in tax to facilitate compliance and assist professionals and their clients with commonly encountered questions. While data analytics has received a lot of attention, Artificial Intelligence in tax is a relatively new phenomenon.”

Not everyone is jumping on this bandwagon. Kuzniacki wrote on January 25, 2019: “All of the features that are indispensable to lawyers … have until recently also appeared to be extremely resistant to AI. That is to say, the complexity, uncertainty and dynamic nature of legal reasoning have presented significant barriers to the development of commercial AI applications. On the supply side, moreover, developing an AI program applicable to law is very time consuming and extremely expensive. On the demand side, the cost-effectiveness of a stand-alone computer equipped with the traditional applications for the legal professions (e.g. statutory and case law databases, commentaries to laws and cases) far exceed the potential gains of investing in the development of an AI program capable of applying the law.”

Kuzniacki addressed the “Present and the Future: Hopefully just Augmenting but Never Replacing of Tax Lawyers” accordingly: “In September 2013, Frey and Osborne from the University of Oxford published the results of their research on the probability of computerisation (i.e. job automation by means of computer-controlled equipment) in 702 detailed occupations in the US, including legal professionals (lawyers). To estimate probability they used a novel methodology using a Gaussian process classifier, which appears in many contexts such as statistics, probability theory and machine learning. Pivotal to the current study is their finding that lawyers are generally not fully computerisable, or, so to say, they belong to the group of least-computerisable occupations with a probability of only 3.5 per cent of being more or less replaced by automatized computer systems.

“By comparison, tax examiners and collectors, and revenue agents were classified as fully computerisable with 93 per cent probability, which is more than for taxi drivers (89 per cent) or parking lot attendants (87 per cent). Recreational therapists, in turn, were classified as the least-computerisable occupation.

“More specifically, Frey and Osborne observed that occupations that involve complex perception, creative intelligence tasks, and social intelligence tasks (i.e. cognitive non-routine tasks) are likely to be supplemented rather than substituted by AI “over the next decade or two.” Their research confirms current ideas that AI is best suited to play a complementary role in tasks performed by lawyers.”

In other words, other tax professionals are likely to use AI — or be replaced. How then could lawyers avoid the prospect? We turn to Canadian lawyers to ask them for their view on the capacity of AI to interpret, let alone analyze, Tax Law.

According to Robert Kreklewetz of Millar Kreklewetz LLP in Toronto: “The potential constraints of AI appear to be its linear thinking and blindly following past precedents in conducting its analysis and generating its predictions. While it might be able to identify the relevant past cases, really good tax lawyers are sought after for their non-linear (lateral) thinking and their ability to argue for results which might even be at odds with past precedents

“Clients generally don’t come to us to determine what the right answer is, they come to us to figure out how to get to the answer they want. That involves a whole host of skills including understanding the policy and intent of the legislation, understanding different ways to achieve different results, and a whole lot of lateral thinking. I’m not sure the AI, as we presently understand it, can replicate or predict that.”

His partner, John Bassindale, meanwhile says, “I wonder what real application this AI has to complex high-end tax litigation. I expect AI’s reliance on past cases means it would be unable to make an accurate prediction where no prior case law exists (as is often the case in high-end tax litigation). While I can understand why the Department of Justice might want to use this AI to evaluate their high volume of simpler tax cases, I wonder how they are using these predictions? For example, if the software predicts the Department of Justice will lose a case, do they settle that case 100 per cent in favour of the taxpayer, or do they try to settle the case for 70 per cent or 80 per cent?”

Blake, Cassels & Graydon LLP’s Mark Tonkovich adds context to this discussion: “Media reports last year indicated that a segment of the Department of Justice’s tax personnel was taking part in an AI pilot project: using predictive software to help analyze tax cases. It’s not clear how the government is using the new software — to supplement traditional case law research, screen new tax appeals, improve efficiencies in settlement negotiations, decide what facts to focus on in court, or for some other purpose. But as today’s software continues to grow in sophistication and expands to cover more hotly disputed tax issues, we can expect both the public and the private sectors to increase their use of the new technology. 

“Whether there will soon come a day when some tax appeals are decided primarily by feeding a summary of the case into an AI is hard to say, but knowing that the tax administration is alive to these new tools has upped the stakes for staying on top of the developing technology.”

David Chodikoff, a partner and national leader within the tax litigation and customs disputes resolution group at Miller Thomson LLP in Toronto, discussed AI in Tax Law on Canadian Lawyer’s website: “In today’s competitive legal landscape, clients are demanding bulletproof advice at a reasonable cost. Firms that adopt AI-backed legal research software are taking proactive measures to ensure that clients walk away knowing they’ve received the highest quality legal advice without paying exorbitant fees.

“‘At the end of the day, clients care about results and costs,’ says Chodikoff. ‘Clients want excellence of service, problem solving at the highest level, and cost efficiencies. In addition to the obvious time-saving benefit, AI-based legal research tools offer lawyers a quick way to access data-backed support for their professional hunches. 

“‘Artificial intelligence adds a dimension to your thinking,’ says Chodikoff. ‘The analysis identifies insights that you might not have thought of and may lead you to a case that adds to your approach to a particular issue. That’s invaluable — you can’t put a price on that.’

“The AI tool that Chodikoff mentions is Tax Foresight, a joint effort between Blue J Legal and Thomson Reuters. The software applies AI to all relevant past judicial decisions in an effort to help lawyers and other tax professionals determine the strength of their position on issues like real estate, taxable benefits, carrying on business, worker classification, and many others. The software also has an advanced search function that finds cases by specific factors, rather than by keyword or boolean searches. 

“‘With Tax Foresight, you can find a case that helps with an argument and you never know if that’s going to be the winning argument before a court,’ says Chodikoff.

“On top of making lawyers more efficient, Chodikoff says that Tax Foresight is comprehensive and acts as ‘additional blanket coverage’ by considering every relevant case in the selected area of tax law.

“‘Prior to the availability of Tax Foresight and the rapid analysis of information that it enables, I’d either conduct my own research or rely on the assistance of an associate or student of law to conduct that research.’”