AI and the R&D Tax Credit: Which Activities May Qualify?

Artificial intelligence (AI) is rapidly transforming industries by helping businesses automate repetitive processes, improve decision-making and develop innovative products and services. As organizations invest significant resources into AI initiatives, many are asking whether these activities may qualify for the federal Research & Development (R&D) Tax Credit under Internal Revenue Code (IRC) Section 41.

The answer is often yes, but qualification depends on the nature of the activities being performed rather than the fact that the project involves artificial intelligence. Under IRC §41, AI projects must satisfy the same requirements that apply to any other qualifying research activity.

The Four-Part Test Applies to AI Projects

To qualify for the R&D tax credit, activities generally must:

Many AI projects may naturally involve these elements. Development teams frequently evaluate multiple model architectures, training methods, datasets and optimization techniques to determine which approach achieves the desired outcome. This trial-and-error process may satisfy the experimentation requirement under Section 41.

AI Activities That May Qualify for the R&D Tax Credit

Potentially qualifying activities may include, but are not limited to the following:

The key factor is whether the team is attempting to resolve technical uncertainty through a process of experimentation rather than simply implementing, configuring or integrating existing technology. Activities are generally more likely to qualify when they involve the development of new, internally created technological capabilities, algorithms, models, software or processes that did not previously exist within the organization. In contrast, the use, customization, deployment or routine application of commercially available AI tools, software platforms or established technologies, without the development of new underlying technology, generally would not be considered qualifying research activities.

Qualified Research Expenses

If an AI project qualifies under Section 41, certain related expenses may qualify, typically including:

For many AI developers, employee wages for software engineers, data scientists and technical managers represent the largest component of a research credit claim.

Documenting AI Research Activities

As the IRS continues to increase its focus on research credit substantiation, documentation has become increasingly important. Taxpayers should maintain records demonstrating the technical uncertainty encountered, alternatives evaluated, testing performed, and resources devoted to the project. Relevant documentation may include the following:

Strong documentation can help establish how the project satisfies the requirements of Section 41 and support the inclusion of associated qualified research expenses.

Final Thoughts

Artificial intelligence projects may present significant R&D tax credit opportunities when development activities involve technical uncertainty and a process of experimentation. While simply using AI does not create eligibility, organizations developing or improving AI technologies should evaluate whether their activities satisfy the requirements of Section 41.

As AI investments continue to grow, companies that properly identify, document and substantiate qualifying activities may be able to generate valuable federal tax credits while continuing to innovate.