As Tax Day approaches, millions of people are using tax software, such as TurboTax, to prepare their tax returns. But what if you make a legal error on your tax return as a result of the tax preparation software? Under current law, the legal liability for the error is directly on you - the taxpayer.
In her new work, Susan Morse proposes to fundamentally change the way regulatory law is enforced. She proposes government-to-robot enforcement. Specifically, Morse argues that an automated law system, which is any machine that produces a legal determination, should be held directly liable for compliance errors made by its users. Therefore, if you use TurboTax to prepare your taxes and you correctly input your facts, but the system produces a return that understates your tax liability, you would not be directly liable for this error. Instead, if the error is discovered, the IRS would pursue enforcement against and impose liabilities directly on TurboTax.
There are several key elements to Morse’s argument. First, the current system creates a problem of negative externalities. It is often the innocent public, rather than the regulated party, who bears the burden of undetected errors. Thus, as Morse concludes, “the development of automated, centralized law systems presents the perfect opportunity to force regulated parties to internalize such negative externalities.” To further ensure that costs and benefits are correctly allocated, Morse persuasively argues that imposing strict liability in this context is both necessary and appropriate. By imposing direct and strict liability on the automated law systems, this cost will flow to the regulated users as the systems increase their prices to account for this additional liability and ensure that it is the regulated users rather than innocent third parties that bear the cost of noncompliance. It may also promote market differentiation as users can choose between different automated law systems based on their risk-tolerance.
Second, under-detection and under-enforcement are major vulnerabilities of the current system, which prevents legal compliance from functioning effectively. Implementing government-to-robot enforcement could help address these problems by targeting the automated law systems, rather than the much larger pool of regulated users. Morse also suggests the use of a damages multiplier to further address the problem of under-enforcement, as well as improve the internalization of negative externalities.
Finally, the work concludes by recognizing that a government-to-robot enforcement system is not without its shortcomings. For instance, a serious concern with this type of system is that the makers of automated law systems may persuade the government to create guidance that favors their users. This “capture” of government is undesirable in that it gives too much power to influential interest groups to sway the direction of the law. Similarly, another significant concern is reverse capture, or the risk that the automated law system will only implement conservative, government views into the system. Other disadvantages include the decline of individual disputes with the government and, as with any new system, the creation of some winners and some losers.
In sum, this work makes an important and valuable contribution to the scholarship on compliance and enforcement methods and should be of interest to both tax scholars and the legal community more broadly. As we move further into the new automation era, these new technologies raise novel issues but also provide us with a unique opportunity to revise our current approach to legal and regulatory issues. Although work remains to be done to address some of the concerns raised by a system that imposes direct liability on automated law systems, Morse’s thought-provoking article encourages a serious consideration of the merits of government-to-robot enforcement.