TAX-I: your virtual legal research assistant
AI case 8/16 of applied artificial intelligence
Whether searching for relevant case law, analysing rulings, or assessing whether a tax case is likely to be successful, tax lawyers have a lot on their plate, and the result is by no means always accurate. What if you could automate this preliminary legal work?
It would enable a tax case to not only be settled more quickly, but also more efficiently and accurately. That was the idea behind TAX-I, a virtual legal research assistant developed by Deloitte, which is able to use artificial intelligence to analyse thousands of tax cases of the European Court of Justice, relate them to similar cases, summarise them, and even predict how a court would rule in a case.
AI email alert
Receive the latest AI casesSign-up
Categorising and learning
“EU Member States are required by law to publish their rulings and to make them easily accessible to third parties,” says Marc Derksen, Consultant in Data Analytics & A.I. at Deloitte Indirect Tax, and a member of the team that developed TAX-I. “So there are bulk quantities of data available.” Working in two sprints of six weeks, the TAX-I team examined what they were able to do with it, and it soon turned out that the possibilities were numerous.
First of all, the tool visualises the way in which a new case corresponds to earlier cases. An interactive chart of lines shows how cases relate to one another, and the size of a dot indicates the relevance of the case based on the number of references. The tool also produces a summary of all cases, based on how often sentences or parts of them repeatedly occur in a ruling. “The summaries still aren’t well-written, flowing texts,” remarks Derksen. “We intend to improve that in the future. In any case, the summaries are currently at a level that enables you to assess what the case is about.”
Finally, TAX-I is able to predict how the European Court of Justice is likely to rule in a case, based on facts that the user of the tool can enter him/herself. It uses a machine learning algorithm that is trained to recognise patterns in tax cases and to draw conclusions. All 1153 tax cases of the EU Court of Justice have now been entered into TAX-I. The team also hopes to analyse all cases from individual countries later on. “In the Netherlands alone, that already gives us 2.5 million cases,” says Derksen. “So much data will enable us to use deep learning, and produce more accurate insights.”
The team has been working for a few months now on making TAX-I simpler and more powerful. The analyses of TAX-I are being evaluated in the meantime by Deloitte’s scientific office and in a study by Vrije Universiteit Amsterdam and Tilburg University.
In the long term, TAX-I is intended to become accessible for all Tax & Legal consultants in Europe. “It will enable them to save time and improve the quality of their services,” asserts Derksen. But the current prototype has already proven its worth. “I regularly speak to colleagues who found related cases using TAX-I that they themselves had missed. That is a valuable achievement by itself.”
*) This case is part of the series of 16 Artificial Intelligence projects from Deloitte. Other cases in the series are in random order:
- TAX-I: A virtual legal research assistant
- AI Benchmark
- SONAR: Find labelling errors in databases
- Transaction detector with regard to the Dutch work cost regulations
- GRAPA: assistance with risk strategies
- Chatbot as a handy search tool for the online technical library
- Argus: an eye for detail
- PostNL: optimising delivery times
- Virtual assistants: beyond the hype
- HR agent Edgy: the future of Human Resources
- Using machine learning to assess risks for insurance policies
- Predicting payment behaviour
- DocQMiner: contract analysis performed in no time at all
- Combating welfare fraud with machine learning
- Using machine learning and network analytics to search for a needle in a haystack
- Clustering unstructured information in BrainSpace
Sign up for the email alert to get all cases through email.