Last month, a colleague reflected on the accelerating train of international taxation, but that’s not the only area that’s accelerating. It’s a smart “robot” that can do research, type, correct text, reply to an email, change style and tone, retrieve large files, spit out a brief extract, compare, analyze data... . All it takes is the right “prompt”.
Remember how, in our “tax childhood”, we used to run around with binders, always printing and rewriting, underlining with colored markers, spending weeks in data-rooms, downloading and reading tons of documents? The last decade or two of digitization and automation has been like switching from a local train to a bullet train. However, in my opinion the trajectory has now sped up exponentially and the word “acceleration” has taken on a whole new dimension.
The potential of GenAI is huge. I asked my new friend EYQ if he would eventually replace me. In his well-mannered (so far) response, he listed the areas where he’s confident (automation of routine tasks, calculations, analysis, monitoring legislative changes) and the areas where he says he can’t (yet) replace me (interpretations, conclusions, recommendations, strategic thinking, a “personal touch and trust”).
Did you find yourself getting carried away for a moment, envisioning yourself thinking strategically on a beach, cocktail in hand, fleetingly supervising the robot working on the lounge chair beside you, finalizing a VAT return, explaining the reasons for deferred tax movements in quarterly reporting, drafting a response to a call for an audit report, collecting and sorting data for Pillar 2 and preparing a brief summary of the main points of an amendment and the related explanatory report?
Let’s just remember to look at the full picture, one in which the tax administration will eventually be similarly equipped. While, logically, there will be a lag between the private and public spheres, the tax administration is also stepping on the gas, both legislatively, in all international initiatives and information exchanges, and in its analytical and control activities. The vast difference in the quality of tax controls today and twenty years ago probably needs no mention. So presumably there will be a robot working on the lounge chair on your other side analyzing your accounting ledger and all your documents and preparing a crisp challenge concerning what it doesn’t like and what it would like to assess.
Another area that fascinates me about AI and robots is whether and how they will be taxed. The first in-depth discussions on this topic were held back in 2017. Considerations started with the simple idea of taxing robots as employees, which was primarily based on concerns about the loss of tax revenue when human labor is replaced by automation. People will need unemployment benefits, robots won’t retire or take maternity leave, won’t give birth. A fiscal dream. But difficult to implement; we aren’t talking about replacement anymore, especially with GenAI, and it wouldn’t be quantifiable, anyway.
Other proposals have tried to tax corporations that use AI more. Does this remind you of the sectoral tax, the windfall tax? Also (logically) off the table – all efforts are now sunk into supporting AI and automation, even in the form of subsidies and R&D incentives, so it doesn’t resonate back in the form of higher taxation. Some studies also mention the potential for some form of VAT taxation, but only very marginally, as it lacks both logic and implementation potential.
Finally, I asked EYQ how he sees all of this. We mostly agreed. Taxation is tempting, motivated by fiscal needs, but currently counterproductive as the AI field needs to be further developed and invested in. Also, practical implementation will be difficult.
Enjoy the newsletter while the authors of the articles are still human... .