QRE that could qualify for the R&D tax credit include amounts paid for:
- Salaries
- Supplies
- Contract research
- Cloud computing costs
Salaries paid to employees who engage in qualified activities are frequently one of the largest components in an R&D tax credit claim. Generally, employees that evaluate technical alternatives or use simulations, models, or tests to resolve technical uncertainties will qualify for inclusion in the credit calculation.
For example, a company with employees in roles such as machine learning engineers, software engineers, data scientists, business analysts, or statisticians that participate in developing and testing AI technologies can include costs for those employees in the credit calculation.
Other individuals, including U.S.-based contractors, who perform similar activities or supervise and support staff in resolving technical uncertainties, may have also includable expenses.
Many companies experimenting with AI use cloud computing services and other computing technologies like continuous integration, continuous delivery, and containerization. Expenses for use of cloud-based computing services in experimentation can also be included. As an example, a company with total expenses of $1,000,000 for software engineers, U.S.-based contractors, and cloud computing could generate a $100,000 credit. If the company is a startup, it may be eligible to offset payroll taxes using the credit.
State opportunities
In addition to the federal benefit, many states have legislation that mirrors the federal rules, further increasing available credits.
Claiming the R&D tax credit
To claim the R&D tax credit, taxpayers must evaluate and document research activities contemporaneously to establish expenses paid for each qualified research activity.
Changes to rules for research and experimentation expenses
Recent changes to Section 174, which governs treatment of research and experimentation expenses, including software development, have an adverse impact on the ability of companies performing development in the AI space to mitigate development expenses.
Specifically, the 2017 Tax Cut and Jobs Act requires software development expenses to be capitalized and amortized over a multiyear period beginning in 2022. The net effect is that only a portion of development expenses can be deducted in the year incurred.
Section 41 permits taxpayers to claim an R&D tax credit for software development expenses, thereby partially offsetting the loss of the Section 174 deduction.
At the federal level, the R&D credit carries forward for up to 20 years, so taxpayers without immediate use may still benefit from claiming a credit and carrying it forwards for future use.
Recovery startup ERC credit
The ERC credit is another tax credit of potential interest to companies that use AI technology.
What is the ERC credit?
As part of the COVID-19 pandemic relief, Congress created a refundable payroll tax credit for companies with businesses that
were adversely impacted. Congress also included a specific benefit for companies that were launched in the midst of the pandemic, though this benefit is often overlooked. Businesses may qualify for the ERC Credit under the general qualification criteria for periods beginning as early as March 13, 2020. New start-ups can uniquely qualify under the recovery start-up criteria if they don’t meet the other general qualifications.
The ERC credit provides a payroll tax credit of up to $100,000 for recent startups. This credit may be used in conjunction with an R&D payroll tax credit election of up to $500,000 annually to offset costs incurred for R&D involving AI technologies. The credit applies to companies in technology related fields, as well as other industries that had salary expenses in Q3 or Q4 of 2021.
Qualifying for the ERC credit
A recovery startup business generally is an employer that meets the following criteria:
- Began a trade or business after Feb. 15, 2020
- Has average annual gross receipts of less than or equal to $1 million for the three-taxable-year period ending with the tax year that precedes the calendar quarter when the credit is claimed
- Not eligible under any other qualification criteria
Companies that began a trade or business after Feb. 15, 2020, may be eligible for $100,000 in refundable payroll tax credits for costs incurred in Q3 and Q4 of 2021.
Many AI start-ups that ramped up hiring in 2020 or 2021 may qualify for this incentive, as the technology became increasingly popular during the transition to remote work. The IRS is increasing scrutiny of refundable payroll tax credits due to many fraudulent claims, so it’s important to seek out a provider who’s knowledgeable in this tax area.
How different industries explore AI development
Some examples of how various industries use AI that may be eligible for the R&D tax Credit include:
Healthcare
A lot of software companies are using AI to perform statistical analysis on large quantities of healthcare data. Example data sets include demographic information, patient profiles, and treatment history. AI can be used to identify statistical correlations in the data across patient populations, thereby improving delivery of healthcare services, identifying at-risk patients, and reducing costs.
Engineering and materials science
Creating data sets about existing materials can help develop new ones, where AI can then be used for designs that optimize competing parameters, such as strength, weight, and cost. This has applications in fields like aerospace and automotive, where shedding weight provides knock-on benefits.
Pharmaceutical
The pharmaceutical industry uses AI to expedite protein folding to identify the mechanism by which proteins interact with various chemical compounds as a step in identifying pharmaceutical candidates.
Game development
AI can be used for creating new game features, like procedurally generated environments and new types of game assets. This is a labor-intensive process involving interactions between various conceptual and technical artists, as well as software engineers.
Cybersecurity
Many companies use AI to develop baseline profiles of normal network traffic. When networking traffic begins to deviate from historical norms, the AI-driven security systems can alert system administrators and identify the type of anomalous activity, as well as the system involved.
Insurance
AI has made headlines for its ability to respond to customer queries on diverse subject matters. Massive data sets are used to generate responses to customer queries, which reduces man hours needed for customer service functions.
