Will The Rise Of Artificial Intelligence Technology Threaten Social Security’s Solvency?

// The newest tech fad is artificial intelligence (AI). And the face of the technology is ChatGPT which is a large language model that helps people brainstorm, write articles, translate text, conduct research, generate ideas, cheat on exams, or simply just have someone to talk to. With its ability to “speak” in simple, conversational language, ChatGPT is a glimpse of what AI technology is capable of.

AI can generally process certain types of information faster than humans.  Initially, the technology will assist humans with their work. But eventually, AI will slowly replace some of these workers altogether. These workers pay payroll taxes which fund Social Security, Medicare, and state programs. If these workers disappear, then the tax base declines and these safety net programs are in jeopardy. How can the government prepare for this?

googletag.cmd.push( function() { // Display ad. googletag.display( "div-id-for-top-300x250" ); });

For clarification, let’s compare the the tax treatment of human employees and machines. Both the employer and employee equally pay a portion of an employee’s payroll tax which is based on their income. Employers also pay similar employment taxes to the state to fund its unemployment and disability programs. Self-employed people similarly pay self-employment taxes on their income. For 2023, only the first $160,200 of wages or self-employment income is subject to Social Security tax every year.

Also, employers must generally file quarterly employment tax returns with the IRS and the state.

Conversely, if a business purchases equipment that uses AI technology, it is not subject to federal and state employment taxes and return filing requirements. This is because AI is not considered an employee. At least not yet. Instead, the business can opt to take a tax deduction for the full purchase through a Section 179 deduction. Or it can choose to deduct a portion of the cost over a number of years through depreciation. The cost of routine repairs is fully deductible. The cost of major equipment upgrades can be either fully deducted or depreciated.

googletag.cmd.push( function() { // Display ad. googletag.display( "div-id-for-middle-300x250" ); }); googletag.cmd.push( function() { // Display ad. googletag.display( "div-id-for-storycontent-440x100" ); });

Due to its nature, AI will not demand things like a living wage or remote work. Nor will it unionize, go on strike, or file employment-related lawsuits.

This is nothing new as businesses have in the part purchased equipment to automate certain tasks and make existing work more efficient. Computers and even smartphones have made computational work easier. But recently, AI has been able to perform creative tasks, such as write poems, essays, articles and even scripts. The use of AI is a point of contention in the 2023 Writers Strike. One demand is that AI can only be used as a tool to help research or facilitate script ideas and not to replace writers.

As AI technology improves, the role of humans in the labor force will change. Workers with skills in robotics or computer science will be in demand and are likely to be highly paid. So in the short run, there may not be a dramatic effect on payroll taxes as young people or those seeking a career change adapt. But older people and those who devoted a great deal of time in their current professions might be unable or unwilling to become robot repairmen.

Assuming that AI technology has advanced enough to threaten the solvency of Social Security, unless the system is fundamentally changed, taxes will have to increase. But who should have to pay the lost payroll taxes?

One option is to have the AI manufacturers pay through an excise or supplemental tax. But this would create three problems. First, it will be hard to determine what technology would be considered AI and thus subject to the tax. Will it need to have sentience or its own consciousness and can it give advice or make decisions? If AI is simply defined as a program that reduces or eliminates the need for human labor, then a lot of everyday items like a computer or a smartphone would be subject to this tax.

googletag.cmd.push( function() { // Display ad. googletag.display( "div-id-for-bottom-300x250" ); });

The other problem is that the excise tax will be passed on to customers. This assumes there is little to no competition and economic substitutes.

The final and most significant problem is that taxing AI technology could hinder or even destroy its development in the U.S. Experts in the field will move to other countries to continue their work. This could put the U.S. in an economic and even military disadvantage.

Another option is to tax businesses that use AI to replace workers. One way to do it is to reduce or disallow a business tax deduction for the purchase of AI technology (South Korea did this in 2018). Also, maintenance expenses paid can be subject to payroll taxes similar to how employees are paid. Like the manufacturers, businesses will try to pass on the tax to its customers through increased prices although, if the business has a large amount of competitors, it will have to pay the tax to stay price competitive.

The final option is to force taxpayers to pay more. This can be done by raising the age for eligibility to receive social security benefits or increase the payroll and self-employment tax rates. Obviously, people are not going to like having to pay more and wait longer for benefits. Many are living paycheck to paycheck. A forced tax increase could result in reduced spending on other areas. This can have negative collateral economic effects which can last years before the economy adapts.

While AI-driven job displacement could potentially impact social security funding by altering the composition of the workforce, the overall effect depends on a variety of complex factors. If only something was available to quickly digest all of this complex information into advice in a conversational format.

Steven Chung is a tax attorney in Los Angeles, California. He helps people with basic tax planning and resolve tax disputes. He is also sympathetic to people with large student loans. He can be reached via email at [email protected]. Or you can connect with him on Twitter (@stevenchung) and connect with him on LinkedIn.

Topics

AI Legal Beat, Artificial Intelligence (AI), Government, Social Security, Steven Chung, Tax Law, Technology