The Break-Even Point of Automating Tax Evasion and White Collar Crime Capture

Infect Teh Interwebs

Right now the vast majority of types of tax evasion, tax avoidance and tax fraud are not caught in the developed world. As a result countries lose an enormous amount of revenues and end up with massive budget holes. The tax agencies simply do not have the manpower to go after various forms of white collar fraud, especially in the US where corporations tend to dictate terms and litigate any attempt towards rectification into oblivion.

However – artificial intelligence is advancing at a breakneck speed. Right now we have a doubling rate of AI capacity in terms of months rather than the Moore’s law standard of years. Hence I would regard it as likely and highly desirable that governments invest in developing AI tools for going after corporate crime (tax and otherwise). There is a point where the capacity of AI to create relatively airtight evidence for white collar crime thereby facilitating quicker sentencing. I do not see the capacity of white collar corporate criminal AI keeping up with government AI since AI research tends to be fairly visible in science papers. So governments can look at the state of the art of subterfuge technologies and adjust their attempts accordingly.

It is hard to estimate when the break even point is reached (where investment in these technologies is markedly less than the revenue it engenders) and the very moment we get there libertarians fraudsters in the world will cry their eyes out as governments finally know how to find them and hold them accountable.

A subsidiary of this idea is that we can likewise algorithmize government for effectiveness in policy not many years later, and create government that works for everyone rather than government that works only for lobbyists and their sponsors. Right now we still inhabit an injust era where the proverbial 1% of people can coerce politicians and skew laws in their favour. We may soon anticipate an era where this is no longer possible.

When will it be possible to automate justice, so to speak? My educated guess is within ten years, but no more than 15 years. That’s a monumental conclusion that will have far reaching consequences for many countries world wide, and one that will quickly trickle down to people as more tax revenue generally means better national resources and infrastructure that can be used for the actual population.

If you read this and agree please spread this Op-Ed so as many politicians, AI researchers and civil servants world wide as possible can draw the required conclusions. Clearly we will need some sort of basic income in the next few years, and using AI to give governments more agency is absolutely necessary.