In Brief

To date, “smart contracts” have had only limited use. Nevertheless, they have the power to be transformative, particularly when combined with artificial intelligence (AI). The authors examine the possible combination of these two technologies and the potential opportunities it offers to resolve asset valuations, a perennially thorny issue between taxpayers and the IRS.


In 1997, in lieu of traditional legal advocacy where parties go through a tiresome, expensive, and nettlesome process of negotiation, computer scientist and legal scholar Nick Szabo conceived of the idea of “smart contracts” for those contractual relationships that may be formalized through executable computer code (“View of Formalizing and Securing Relationships on Public Networks,” First Monday, doi: 10.5210/fm.v2i9.548). To date, smart contracts have been used in various narrow settings, most notably in cryptocurrency and in certain insurance contracts. Despite their limited use, they have the promise to potentially revolutionize the tax and accounting worlds (see the Exhibit). They can be employed as a mechanism for taxpayers and the government to reach a consensus as to asset valuations, an area of historic contention that has resulted in myriads of adjudications and audited statements punctuated with caveats.

The analysis below begins by delving into the problematic issue of asset valuation. It is followed by an exploration of the nature of both smart contracts and artificial intelligence (AI) and their capacity to address the endemic asset valuation problem, concluding with a description of the benefits associated with this proposed approach.

Background: Asset Valuations

Since the inception of the income tax in 1913, asset valuation issues have been omnipresent. And the reason is simple—by their nature, nonfungible assets have a range of values. Depending upon the taxpayer’s agenda, valuation choices will be made that yield the least possible tax burden.

If, for example, a taxpayer wishes to donate property to a charitable organization, along with a permissible asset valuation range, they will choose the highest one. Conversely, if a taxpayer wishes to gift property to their children, along a permissible asset valuation range, they will choose the lowest one. Securing defensible asset valuations (often from a professional appraiser), however, involve resource and time-intensive exercises.

Given taxpayers’ tax-saving agenda, on many occasions, the IRS disputes the proffered valuation figures. Often, what then transpires are long litigation battles between competing experts regarding the accuracy of the asset valuation. With little or no technical valuation training to resolve these issues, the courts are then tasked to resolve the parties’ differences, frequently averaging the dichotomous choices presented.

The result of this dynamic is predictable; it causes the IRS and taxpayers to take even more extreme valuation positions in the expectation that doing so will, from their perspective, ultimately yield more favorable valuation outcomes.


Potential Smart Contract Applications

 Debt Covenants; Smart contracts can automatically monitor and operate a loan covenant. The conditional terms of a loan covenant are encoded into a smart contract that is designed to oversee a company's financial conditions. If a violation of the term is detected, a penalty may be imposed automatically (e.g., an increase in the interest rate or termination of a debt agreement). Government Contracting; During the government contracting bidding process, procedural rules based upon a defined set of criteria could be encoded into a smart contract. A bidder would then be automatically chosen based on the encoded smart contract, increasing transparency and fairness. Supply Chain; Using smart contracts and capitalizing upon blockchain, bills of lading can be transferred both electronically and securely to log ownership changes and automatically transfer payments to ports upon arrival. Employment Contracts; Standardized wage systems could be embedded into smart contracts that could not be easily changed or overwritten and, due to their uniformity, would reduce claims of wage discrimination. Healthcare; Smart contracts can provide access control for medical records that are stored on the blockchain. Such records could be securely protected by cryptography, thus not visible to the public. To prevent manipulation, only related parties such as medical staff and patients would be allowed access to records. Accounting Records; Based on preset accounting rules, once a transaction occurs, smart contracts could automatically record accounting entries and associated information. Ownership Transfer; Smart contracts can be used to verify and transfer ownership of property such as real estate; furthermore, once there is a meeting of the minds, smart contracts can automatically execute the exchange of property title and fund transfers into an escrow account.

The Intersection of Smart Contracts and AI

Over the past century, valuation battles between the IRS and taxpayers have repeated ad infinitum. The following fact scenario is commonplace: a taxpayer reports an asset’s value, the IRS audits the taxpayer’s tax return and challenges the asset value’s bona fides, litigation ensues, and the courts are called upon to resolve the differences.

Throughout this same period, a technological revolution has unfolded that has completely transformed the ways people conduct their personal and business affairs. For purposes of this analysis, there are two particular technological threads that, when woven together, harbor the potential to alter the traditional dynamics between the IRS and taxpayers. They are smart contracts and AI.

Smart contracts are computer programs designed to automatically perform certain tasks based on pre-determined actions and triggers. As such, these programs are designed to be hermeneutic (i.e., self-executing, self-contained); thus, once executed, intervention by the participating parties, lawyers, or courts may be very limited, if not altogether absent. The absence of these possible frictions makes them “smart” and attractive relative to traditional contracts. One analogy to a smart contract operation is a vending machine, which is triggered by the insertion of coins, bills, or a credit card, followed by the selection of an item, and then concluded by the release of the corresponding merchandise.

In the past decade, a novel technology that has been commonly used to augment and protect the execution of smart contracts is blockchain. Blockchain is a distributed ledger system that utilizes a network of computers that jointly manage a system and prevents illicit ledger actions. Utilizing the safeguards inherent in blockchain technology, smart contract integrity is nearly impenetrable to malicious changes.

AI broadly describes several functions that can now be accomplished through the use of technology without requiring human intervention. Familiar examples of AI include self-driving cars, robot vacuums, and virtual assistants. A subset of AI is machine learning, which involves a process by which sophisticated computers “learn”—often through a trial-and-error process—rather than being programmed.

Over the last decade, one area of AI that has a track record of proven success has been in the asset valuation realm. Asset valuations, which used to take weeks or months to craft at great cost, can now be produced instantaneously by AI at little or no cost. The results are often more accurate than those prepared by their human counterparts.

Given the functionality of smart contracts and the success AI has had in its capacity to value assets, there is no reason that the strengths of these two technologies cannot synergistically be combined to capitalize upon the virtues of each. The proposal below attempts to do precisely that.


The proposal at hand is straightforward: In cases where taxpayers need to value an asset (e.g., contributing an office building to a charitable organization), they would enter into a smart contract with the IRS. The terms of this contract would be uniform and self-executing for both parties, and would rely exclusively upon AI for asset valuation purposes. A blockchain would manage the contract and ensure that its actions strictly follow the pre-set agreement between the taxpayers and the IRS.

This process would unfold in a series of three steps and be deployed on a blockchain platform.

Step 1.

Computer software would prod the taxpayer to supply critical information about the asset in question (e.g., a building’s location, square footage, date constructed). If adequate information were supplied to meet the AI’s needs to make an accurate valuation estimate, the process would continue to the next step. If the supplied information proved inadequate, the smart contract would automatically solicit additional information (e.g., average monthly net rental income generated or any amenities offered by the building). Once this additional information was supplied, the AI would yield a valuation determination as part of the process. If the AI still lacked sufficient information (e.g., comparables), it would inform the parties that a traditional asset appraisal was in order and that the smart contract was null and void.

Step 2.

The IRS or taxpayer could then accept or reject the posited value. If either party rejected the value determination, the AI would automatically generate a series of additional questions for the taxpayer to complete (e.g., average monthly utility costs), possibly resulting in a recalibrated asset value. Once again, the IRS and taxpayers would be in a position to accept or reject the more refined asset valuation.

Step 3.

If either the IRS or taxpayer rejected the asset’s projected fair market value, the parties—as part of the smart contract—would commit to arbitrate their differences or switch to traditional asset appraisal if the valuation differences remained significantly large. Once both parties agreed on the asset’s value, the smart contract would be deployed on a blockchain platform, and the taxpayer would commit to report this value and the IRS to accept it. A smart contract violation could result in the taxpayer being penalized or the IRS being obligated to pay restitution. Execution of penalty and restitution would be enforced by the smart contract and overseen by all computers in the blockchain.

Once bound by the terms of this three-step process, the IRS and taxpayers would have to act in conformity. Hopefully, each party would be pleased with the smart contract’s outcome; however, if one party or both were not, they could not contest or challenge the outcome.

Benefits of the Smart Contract Proposal

The adoption of this proposal would offer a myriad of benefits. For starters, many empirical studies indicate that the valuations produced by AI and related technologies are often more accurate than those fashioned by human experts in the appraisal field. Also, such asset valuations are not limited to real estate; AI has a proven track record that includes producing asset valuations for artwork, closely held businesses, and other unique property (e.g., rare coins, jewelry). See the Sidebar for a sampling of the research in this area.

Beyond the asset valuation accuracy that AI produces, the time and expense associated with the production of such appraisals are a mere fraction of the cost of similar appraisals made at the hands of valuation experts. Asset appraisals ordinarily cost thousands of dollars and can take weeks—even months—to prepare and secure. The exact opposite is the case with AI appraisals; depending upon the nature of the property involved, once the model has been developed, it may only require a few keystrokes to run an appraisal. The cost to procure them is often minimal, and the time it takes to secure them can be instantaneous.

Were the IRS and taxpayers to agree to enter into smart contracts that tap into AI, the amount of litigation involving asset valuation disputes would be greatly diminished. Each party would have no incentive to take extreme values (as is currently the case). Instead, via smart contracts and AI, an asset’s value would be distilled down to one number that would have the presumption of accuracy. Furthermore, blockchain would largely replace the court’s traditional role in monitoring and supervising contract execution.

‘Smart’ Policy

There is a reason technological advances are commonly referred to as being disruptive: they fundamentally change the way humans interact and perform tasks. The authors’ proposal to combine smart contracts and AI in the asset valuation realm is no exception. If, as a matter of public policy, Congress were to mandate their use, taxpayers and decision-makers would have to shed their deep-seated hesitancies. Doing so would constitute an important incremental step towards meaningful reform.

For Further Reading

Aubry, M., Kräussl, R., Manso, G., & Spaenjers, C., “Machine learning, human experts, and the valuation of real assets,” No. 635. CFS Working Paper Series, 2019.

Baldominos, A., Blanco, I., Moreno, A. J., Iturrarte, R., Bernárdez, Ó., & Afonso, C., “Identifying real estate opportunities using machine learning,” Applied Sciences, vol. 8, no. 11, p. 2321, 2018.

Chaphalkar, N. B., & Sandbhor, S., “Use of artificial intelligence in real property valuation,” International Journal of Engineering and Technology, vol. 5, no. 3, pp. 2334-2337, 2013.

Chen, T. H., Chen, M. Y., & Du, G. T., “The determinants of bitcoin’s price: Utilization of GARCH and machine learning approaches,” Computational Economics, vol. 57, no. 1, pp. 267-280, 2021.

Dai, J., & Vasarhelyi, M. A., “Toward blockchain-based accounting and assurance,” Journal of Information Systems, vol. 31, no. 3, pp. 5-21, 2017.

Giulio, P., “Chainalysis and Wave showcase blockchain fintech products at New York Barclays accelerator, sign deal with Barclays,”, 2015.

Guo, J. Q., Chiang, S. H., Liu, M., Yang, C. C., & Guo, K. Y., “Can machine learning algorithms associated with text mining from internet data improve housing price prediction performance?,” International Journal of Strategic Property Management, vol. 24, no. 5, pp. 300-312, 2020.

Guzun, Eugeniu, “What Is It Worth?,” HedgeNordic, Feb. 11, 2020,

Kim, J., Lyrio, M. V., Dai, J., & Vasarhelyi, M. A., “Block-chain-Enabled Continuous Audit: Implementation of Blockchain-Enabled Smart-Contract with the Integration of Business Process Management,” manuscript in preparation, 2022.

Liu, Y., & Zhang, L., “Cryptocurrency valuation: An explainable AI approach,” arXiv preprint arXiv:2201.12893, 2022.

Schatsky, D., “Getting smart about smart contracts,”, 2018.

Yamaura, Y., Kanemaki, N., & Tsuboshita, Y., “The Resale Price Prediction of Secondhand Jewelry Items Using a Multi-modal Deep Model with Iterative Co-Attention,” arXiv preprint arXiv:1907.00661, 2019.

Yu, Y., Lu, J., Shen, D., & Chen, B, “Research on real estate pricing methods based on data mining and machine learning,” Neural Computing and Applications, vol. 33, no. 9, pp. 3925-3937, 2021.

Jun Dai is an assistant professor at Michigan Technological University, Houghton, Mich.
Jumi Kim is a PhD candidate at Rutgers Business School, Newark and New Brunswick, N.J.
Jay A. Soled, JD, LLM, is a professor at Rutgers Business School.
Miklos Vasarhelyi, PhD, is the KPMG Distinguished Professor of Accounting Information Systems and Director of the Rutgers Accounting Research Center and Continuous Auditing & Reporting Lab (CAR Lab); he is also a member of The CPA Journal Editorial Advisory Board.