SpellCheck: An Educational Device to Improve Children’s Spelling; FDA Regulation of Artificial Intelligence in Medical Software

Beamon, Noah, School of Engineering and Applied Science, University of Virginia
Rogers, Hannah, EN-Engineering and Society, University of Virginia
Powell, Harry, EN-Elec/Computer Engr Dept, University of Virginia

This portfolio consists of two projects. A technical report and a Science Technology and Society (STS) research paper. The technical report contains the design, development, and demonstration of a children’s spelling device to understand how game-based learning impacts children’s ability to learn how to spell English words. The STS research paper employs actor network theory (ANT) to address the Food and Drug Administration (FDA) regulatory framework of artificial intelligence (AI) and machine learning (ML) in medical devices and medical software to explain the continuity of the use of computer aided diagnosis (CAD) software in medicine. Although these projects are unrelated, they comprehensively address regulatory implications and the issue of product affordability. I learned that all industries employing technology must successfully manage regulatory and affordability challenges.

The first concept that is applicable to both projects is the consideration of regulatory implications. Children’s educational devices and medical software applications are both bureaucratically regulated in the U.S. The advancements made within these industries and the progress of these industries comprehensively rely on bureaucratic regulatory frameworks. These regulatory frameworks coincidentally facilitate the continuity of these industries through restrictive and expansive measures which ensure safety and efficacy. In Revealed Bureaucratic Preference: Priorities of the Consumer Product Safety Commission, Thomas states, “the Consumer Product Safety Commission (CPSC) is empowered to reduce or to eliminate consumer exposure to unreasonable hazards from every consumer product not elsewhere regulated by the U.S government” (Thomas). In the development of the children’s spelling device, a combination of U.S regulations and international standards was paramount for ensuring the production of the device met physical safety standards including the CPSC standard 16 CFR addressing the safety
of child toys and the IPC standards for printed circuit board (PCB) design. In Software As a Medical Device: FDA Digital Health Regulation, Deloitte states, “In mid-2017 the FDA release[ed] three new guidance documents—two of which distinguish between device types that are low-risk and, therefore, no longer required to undergo pre-market review, and one which outlines new guidelines for evaluating [Software as a Medical Device (SaMD)] applications” (Deloitte). The FDA designates CAD software as SaMD and encourages CAD software engineers to adhere to a set of established standards when integrating AI and ML in CAD software. Both projects involve regulatory implications that influence product development.

The second concept that is applicable to both projects is affordability. Child educational devices and medical software must be affordable and effective, so all individuals have access to the benefits the technology provides. In Realizing the promise: How can education technology improve learning for all?, Ganimain asks the following questions when discussing electronic white boards: “Will these expensive boards be used in the same way as the old chalkboards? Will providing one device (laptop or tablet) to each learner facilitate access to more and better content, or offer students more opportunities to practice and learn?” (Ganimain). Similar inquiries must be considered for the children’s spelling device; engineers must consider whether the device makes an impact and whether it will be accessible to everyone based on its price. In Who Will Pay for AI?, Chen states, “Development of artificial intelligence (AI) in radiology has been much more rapid than in other specialties in health care. . .U.S. regulatory approval is the initial hurdle for adoption of AI in the United States. A much bigger hurdle for broader adoption of new technology is payment” (Chen). Although the first challenge to implementing software into health care is establishing a regulatory framework, establishing an insurance policy to pay
for it and make it accessible to everyone is also a challenge. Both projects involve devices for which affordability is a factor of consideration.

BS (Bachelor of Science)
SpellCheck, Computer Aided Diagnosis, CAD, FDA

School of Engineering and Applied Science
Bachelor of Science in Computer Engineering
Technical Advisor: Harry Powell
STS Advisor: Hannah Rogers
Technical Team Members: Justin Guo, Rachel Lew, Catlinh Nguyen, Shymbolat Tnaliyev

All rights reserved (no additional license for public reuse)
Issued Date: