Faced with the potential for facial recognition to reveal a person’s true identity, some experts are urging a new approach to understanding the illness.
Faced a patient who was not mentally ill, doctors could be tempted to use the technology to identify the patient, but they could be inadvertently identifying someone who has an illness that is not known to be the real illness, said Dr. Jennifer DeMuth, associate director of the Boston University School of Medicine’s Division of Psychiatric Disorders, Behavioral Disorders and Intellectual Disabilities and chair of the Department of Psychiatry at Boston University.
In fact, she said, many doctors are skeptical about using facial recognition at all.
The technology’s limitations also make it an effective tool in identifying people who are depressed, bipolar or suicidal.
“It is an extremely dangerous tool,” DeM, a physician, said.
The challenge for clinicians, who use the tool for things like screening, treatment and identification of suspected cases of mental illness, is to make sure the information they collect is accurate.
“If you collect information about people who do not fit your diagnoses, it could be misinterpreted and could be interpreted as a diagnosis,” DeDuth said.
“So we need to make our assessment as much as possible from the data that we have.”
The challenge facing researchers is finding ways to build the trust needed to use facial recognition effectively.
In a recent paper published in the Journal of Affective Disorders, Dr. Eric Schmitt and colleagues found that facial recognition could help clinicians better identify individuals with mental illness based on their faces and the features of their skin.
“Face recognition is not foolproof,” said Dr
