In the early 2000’s, it was quite clear that technologies using AI would be created not to replace legal professionals, but to enable them to perform their duties at a higher level.
Demands for transparency and the fight against tax evasion, and even tax avoidance by multinationals and bigwigs, have become increasingly significant in the last fifteen years.
Initially in 2012, the multilateral treaty was amended by a number of countries to “update international tax rules and expedite the implementation of a series of tax treaty measures to reduce the likelihood of tax evasion by multinational companies” specifically changing international and national tax rules and corporate tax plans. Led to the rise.
This instability, combined with the increased oversight of tax authorities and increasingly stringent requirements for transparency, could be a challenge for companies. The challenge is not just the need to comply with tax laws; It is also important to ensure the accuracy, continuity and secure storage of tax data.
The responsibility of tax professionals means that they must provide reliable and accurate advice to their clients in order to maintain their tax position.
The tax administration itself has equipped itself with innovative technologies using AI and “machine learning” or even “data mining” to improve efficiency in the fight against tax evasion and better tax collection.
For example, the U.S. tax administration, to name a few, does not hesitate to consult with taxpayers on social networks, specifically using programs capable of analyzing their public data and identifying a taxpayer as a potential tax evader.
A similar process is used by the Canadian Revenue Agency.
The growing complexity of tax regulations, the need for new transparency, the modernization of tax administration that allows for increased efficiency in taxpayer control, and the emergence of many systems that require the use of IAs to compete between different players in tax advice. Taxation
Using AI in taxation
In recent years, new software has been put on the market for companies, individuals and of course tax advice.
They invite tax experts with a promise to stream siren songs more efficiently, provide valuable information, reduce time-consuming tasks, and increase productivity.
While they also claim to use AI and machine learning to get targeted search results in less time, we believe that they are nothing less than advanced search platforms, part of a line of systems created in the 2000s for legal professionals. (For example LexisNexis).
These systems allow professionals to optimize their search skills in a database but do not form predictive tools to help them make their decisions.
One of the most advanced systems using AI and able to predict tax-based solutions based on predetermined situations can be found at start-up Blue J Legal (“Blue J”), founded by three professors at the University of Toronto. Benjamin Alari, Anthony NibletAnd Albert H. Eun.
Prediction system developed by Blue J.
Blue J’s promise is to use machine learning to predict how the court will decide tax matters.
The idea is to save valuable time for tax experts so that they can respond to this problem with greater accuracy and certainty.
According to Blue J, the developed system will have 4 advantages in tax research and analysis:
It will actually be possible to “measure risk for clients, identify the best tax and business planning strategies, discover blind spots, and identify the most effective litigation strategies.”
Blue J guarantees the reliability of more than 90% system results.
How is this possible?
According to its designers, Blue J uses machine learning technology to create a predictive algorithm that will detect connections between different variables.
Machine learning technology is used to create taxonomy classifications based on court decisions issued by the Canadian judiciary (including the Supreme Court of Canada and the Federal Court of Appeal).
These decisions usually confirm or reverse a specific legal question.
The system is designed to answer with some probability how a court will decide to answer this question based on the set of information provided.
Blue J covers a significant portion of tax issues (taxes of Canadian and American companies, taxes on individuals, international taxes, but to name a few).
We’ve been able to test the system for over a week in terms of Canadian corporate taxation, personal taxation, and international taxation based on a variety of circumstances.
Specifically, we were able to run a simulation related to a natural person’s tax accommodation with a significant presence in Canada and a third country.
To answer this person’s tax residency question, you first need to answer a detailed question about the normal person’s situation (see an excerpt of the question asked in the table to the left of the capture below).
In just a few seconds, the system released a report containing the following:
- The potential consequences of the individual’s tax residency which in this case has become Canada;
- The probability percentage of the result which has been indicated to us as 95% specific result;
- A detailed memorandum explaining the rationale for the results based on the answers to the questions raised;
- A summary of the original court decision used by the system;
- Finally, a list of court decisions based on the answers to our questions that are similar to our case.
Without this technology, it must be recognized that an “old” analysis of many court decisions can take hours for a team of lawyers (which will be billed to the last client).
In addition to efficiency, it is highly unlikely that the reliability of such a search would be similar to that enabled by AI.
Blue J makes it possible to create a memorandum or report using the law firm or tax advisor’s graphic charter if desired.
In other words, tax research (which can also refer to a percentage of reliability) can be sent directly to the end customer with minimal intervention from the service provider.
From our previous observations, can we conclude that Blue J has replaced tax lawyers?
Of course not, but there is no doubt that this system provides significant additional value for tax professionals.
According to Blue J, it is also used by most of the major law firms in North America.
Finally, after our various interviews with teams of companies mentioned above, we realize that the growth of tax tools using AI will continue to accelerate in the coming years, both in the private sector and in the public sector.
It’s hard to give AI a blank check
However, the development of systems using AI is not without raising a certain number of questions that are guided by this doctrine and have been compiled in a recent university study.
Without drawing a complete inventory, it seemed important to us to point out a fundamental problem in AI, which is data.
Significant amounts of data are required for machine learning to be effective.
However, this is a raw material that data companies cannot use freely because taxpayers’ tax data constitutes personal information protected by dedicated law, compared to others in Canada (at the state and federal level). Jurisdiction
At the moment, it is “not clearly established” that developers of innovative tax systems have the right to use their clients’ tax data.
For example, for the development of the Blue J system, its designers used case law data as the main source, which was therefore unprotected and freely accessible to the public.
Transcends the paradigm of the instrument that will replace man
According to a survey conducted on more than 700 jobs by teachers Carl Benedict Frey And Michael A. OsborneThe risk of being replaced by technology in the coming years is very uneven from one sector of activity to another.
Thus, the authors of this study show that lawyers are less likely to be replaced by computers or algorithms (with a 3.5% probability rate).
But not everyone is in the same boat.
In fact, the chances of accountants and auditors being replaced are much more worrying because the study estimates that the proportion is 94%!
The major players in tax advice and auditing (especially the Big Four, not to mention their names) have already diversified their portfolio of activities for some time (especially with strategy advice, taxation, legal services when it is authorized in their jurisdiction) to practice, or start- Up incubator).
However, it must be acknowledged that the core business of this structure is closely involved in auditing and accounting activities, which requires a large number of specialized auditors who run hours of service in their client companies, the value of which can be significant.
So one should not be afraid of being replaced by a tax attorney or a tax consultant by machine, but should consider that he will have new “enhanced intelligence tools, which will help him in his decision making”, which will ultimately improve the quality of services provided. To the client.
Member of the Paris and Luxembourg bars, Remy Slama Cyberjustice Laboratory is a lawyer and paralegal who originally published this article. This article is an edited version for brevity. You can see the original text here. He is currently pursuing a master’s degree in law at the University of Montreal.