The University of Washington has canceled a study that would have videotaped preschoolers using a camera worn by their teacher to help train Artificial Intelligence (AI). Parents were outraged after learning about an opt-out research project that would have recorded their child’s likeness and classroom behavior, as well as the classroom environment.
According to reports from 404 Media and Futurism.com, the university researchers decided to cancel the study because parents “erupted in fury” regarding the study's potential privacy violations. At issue was whether or not the study used an opt-out model, which means children could be recorded until parents take action to stop the recording. In most studies conducted with children, there is typically no recording without the explicit consent of at least one parent.
The Consent Problem
When students are enrolled in a research study they are typically automatically included. However, in this case, all children were subject to recording unless their parent took affirmative steps to opt them out. Researchers who work with children also are expected to provide safeguards related to confidentiality and data protection. A critical part of the research process is ensuring informed consent from both the parent and the child.
In this case, some parents expressed concern for their child's safety; however, neither article provided details as to why parents felt this way. The fact that all children were automatically enrolled and the lack of clarity surrounding how the data collected would be handled appears to have led to widespread outrage among parents.
Get the latest model rankings, product launches, and evaluation insights delivered to your inbox.
Capturing Footage
Teachers participating in this study were to wear a small camera while teaching preschoolers. The camera would create a naturalistic dataset based on classroom activities from the first person point of view. Neither of the articles published explained the type of AI application(s) researchers plan to use this dataset for. Prior to the study being cancelled, an Institutional Review Board (IRB) had reviewed and presumably approved the research protocol.
This decision reflects a growing concern about collecting AI-training data in areas such as preschools. More than ever before researchers face barriers due to the increasing sensitivity of people regarding their private information. It is now clear that when researchers collect data concerning children, there will need to be a shift toward a new paradigm for obtaining parental consent prior to conducting AI-related research in schools.