The world has been evolving rapidly in recent years, and technology has developed with it. One of the most important technological innovations is artificial intelligence (AI), which is becoming a key element in the transformation of social and economic life. One of areas where the potential of AI can be exploited is in the broad field of taxation. Traditional methods of analyzing tax regulations are becoming increasingly time-consuming which is due, among other things, to the highly complex legal and tax system. AI assimilates knowledge through the use of advanced algorithms and machine learning. The proper use of the above, by teaching AI the tax law and how to apply it, can not only significantly optimize tax advisors working time, but also contribute to a better understanding of tax law on the part of entrepreneurs.
Tax ruling request generator
Imagine a program dedicated to creating requests for tax ruling. If its knowledge covered all tax regulations in force in Poland, along with issued interpretations of tax law and administrative court decisions, a very useful tool for the work of every tax office could be created. It would be sufficient to describe the background in detail, mark the provisions we want to refer to and, based on the database, the application would create a preliminary version of the justification for our position, together with suggestions regarding the possible extension of the request to include other provisions. If we add to this the indication of the justification of the suggestions on the basis of the latest administrative court rulings, a program with the potential to make a real revolution in the tax world could be created.
Due Diligence and Tax audit
Artificial intelligence (AI) can be very helpful in due diligence, although the safe use of AI in such activities will be possible if a number of conditions are met, including the absence of risk of, for example, client data leakage. Undoubtedly, a program dedicated to such activities will review and analyze documents identifying irregularities much faster and more efficiently. On the other hand, key issues such as the possibility of leaking sensitive client and contractor data or professional confidentiality of e.g. a tax advisor should not be forgotten. It would be irresponsible to use a publicly available artificial intelligence model for such an examination, e.g. due to lack of information on where and how the data uploaded to the AI would be processed (one should bear in mind the legal liability, which in case of data leakage would be on the side of the tax advisor). The emergence of a dedicated due diligence program is likely in the future, although the ‘usability’ of the program in the context of the above-mentioned example risks should be assessed in terms of e.g. the security of the storage of the received data (e.g. by having an internal server and not allowing the program to use the Internet), adequate encryption and liability issues (which in case of e.g. data leakage should be on the side of the program provider). In addition, the issue of assimilation of new data by such a program should not be forgotten – won’t the ability of the law firm’s contractor’s program to teach the program using sensitive data of the company under investigation constitute a breach of the sensitivity of that data?
Due diligence in itself is ‘within the reach’ of the capabilities of AI models, while without legal regulation of the above issues or without dedicated solutions for the legal and tax industry prepared by distributors/developers that will protect law firms in case of establishing cooperation in the form of purchase of such a program or a license to it, the success of such a solution may be doubtful. The above-mentioned risks are only part of the problems to be solved, while in the case of a properly developed model, which will not generate, among others, the above-mentioned risks, automation and reduction of due diligence time may prove possible.
Artificial intelligence may also prove helpful in the case of tax audits, although here the human factor will be more necessary. AI will only analyze ‘dry’ data, as there is no way to upload ethics into the program. The program will not be able to fully assess whether an act bore the hallmarks of a negligible social harm of an act, which can be crucial in many cases. Of course, it is possible to prepare a database for the AI dealing with audits, which will be able to learn from it the definition of the concept of, for example, negligible social harm, but it is difficult to determine at this point whether the AI’s assessment of the situation will take into account morality in the broadest sense. In addition, the aforementioned risk areas cannot be forgotten.
AI – opportunity or threat?
The development of artificial intelligence has the potential to bring many improvements to the daily work of tax practitioners as well as entrepreneurs. Many processes will be able to be automated, although at the same time it should be remembered that the tax advisory profession also involves a number of activities that AI will not so easily replace. Among other things, a tax advisor needs to identify client’s needs and opportunities, respond dynamically to their current difficulties and translate the results of analysis into practical recommendations. In addition, the question of responsibility for the tax advice given remains to be resolved. As mentioned, artificial intelligence can substitute the technical steps of pre-analyzing documents and proposing answers to the questions asked, but it will not be able to take responsibility for its actions. It is hard to imagine a situation in which a client sues the program for bad advice. In the event of any damage resulting from an analysis incorrectly prepared by the program (because, for example, AI fails to take into account a specific provision or a line of case law on a given issue, making the analysis incorrect), the client would have no possibility of compensating for this damage in court. As a result, AI solutions are likely to be a new tool for tax advisers, which will at the same time influence a partial redefinition of their profession.
