Blab AI
University of Pennsylvania AI Tool Achieves 80-85% Accuracy in Autism Screening for Children
Researchers at the University of Pennsylvania (UPenn) have reported that a new artificial‑intelligence system, called CAMI (Computerized Assessment of Motor Imitation), can identify autism spectrum disorder (ASD) in children with an accuracy of 80‑85%. The study, published in a press release by the university and reported by local media on August 9, 2026, involved 183 children aged 6 to 13 who were asked to imitate simple movements shown on a screen.
CAMI works by analyzing a short video of a child’s movements and comparing them to the target actions. The algorithm focuses on the quality of the imitation, a behavioral marker that has been linked to autism, and deliberately ignores environmental variables such as camera angle or lighting. According to the report, the system assigns a score that indicates whether the child’s imitation pattern falls within the range typically seen in children with ASD.
The study’s accuracy figures—80 % to 85 %—were derived from a clinical screening test. While the article does not provide sensitivity or specificity values, the reported range suggests that CAMI performs comparably to existing diagnostic tools that rely on behavioral observation, developmental history, and parent questionnaires.
Professor Rene Vidal, who led the research team, said that CAMI offers a more detailed assessment of motor behavior than current standardized tests. “This is a much more fine‑grained motor assessment than is usually done in the current standardized test,” he told reporters. He added that the team envisions the tool becoming an automated component of a broader diagnostic process.
The researchers emphasized that CAMI is not intended to replace clinicians. “The goal of CAMI is not to replace physicians, but to provide clinics, schools and families with a faster diagnostic tool,” the report states.
The study was conducted in collaboration with Penn Engineering and the Kennedy Krieger Institute, a leading research center for neurodevelopmental disorders. The AI model is part of a broader line of CAMI systems, including CAMI‑2D for video data and CAMI‑3D for motion‑capture data, which have been described in academic papers on arXiv and OpenReview.
Future plans for the project include expanding testing to a larger and more diverse population, including adults, and improving the system’s accuracy. The team also aims to integrate CAMI into clinical workflows and to evaluate its performance in real‑world settings.
The development of CAMI comes amid growing demand for efficient autism screening tools. Current diagnostic procedures can take months to complete and require specialists who are in short supply. An objective, video‑based AI system could reduce waiting times and provide earlier intervention opportunities.
The research was reported by several outlets, including the Philadelphia Weekly, Star News Philly, and 6ABC. The article also references a broader discussion about AI in healthcare, noting that machine‑learning models are increasingly used for disease diagnosis and clinical decision support.
In summary, the UPenn team’s CAMI tool demonstrates promising accuracy in identifying autism in school‑aged children using a simple video test. While the system is not yet a standalone diagnostic, it represents a step toward faster, more objective screening methods that could complement existing clinical practices.
The next phase of the project will focus on larger population studies, adult testing, and further refinement of the algorithm’s performance. The research team has not yet announced plans for regulatory approval or commercial deployment, but the findings suggest potential for broader adoption in educational and clinical settings.
As AI continues to permeate healthcare, tools like CAMI illustrate how machine‑learning models can augment traditional assessment methods, potentially improving access to early diagnosis and intervention for children with autism.
CAMI works by analyzing a short video of a child’s movements and comparing them to the target actions. The algorithm focuses on the quality of the imitation, a behavioral marker that has been linked to autism, and deliberately ignores environmental variables such as camera angle or lighting. According to the report, the system assigns a score that indicates whether the child’s imitation pattern falls within the range typically seen in children with ASD.
The study’s accuracy figures—80 % to 85 %—were derived from a clinical screening test. While the article does not provide sensitivity or specificity values, the reported range suggests that CAMI performs comparably to existing diagnostic tools that rely on behavioral observation, developmental history, and parent questionnaires.
Professor Rene Vidal, who led the research team, said that CAMI offers a more detailed assessment of motor behavior than current standardized tests. “This is a much more fine‑grained motor assessment than is usually done in the current standardized test,” he told reporters. He added that the team envisions the tool becoming an automated component of a broader diagnostic process.
The researchers emphasized that CAMI is not intended to replace clinicians. “The goal of CAMI is not to replace physicians, but to provide clinics, schools and families with a faster diagnostic tool,” the report states.
The study was conducted in collaboration with Penn Engineering and the Kennedy Krieger Institute, a leading research center for neurodevelopmental disorders. The AI model is part of a broader line of CAMI systems, including CAMI‑2D for video data and CAMI‑3D for motion‑capture data, which have been described in academic papers on arXiv and OpenReview.
Future plans for the project include expanding testing to a larger and more diverse population, including adults, and improving the system’s accuracy. The team also aims to integrate CAMI into clinical workflows and to evaluate its performance in real‑world settings.
The development of CAMI comes amid growing demand for efficient autism screening tools. Current diagnostic procedures can take months to complete and require specialists who are in short supply. An objective, video‑based AI system could reduce waiting times and provide earlier intervention opportunities.
The research was reported by several outlets, including the Philadelphia Weekly, Star News Philly, and 6ABC. The article also references a broader discussion about AI in healthcare, noting that machine‑learning models are increasingly used for disease diagnosis and clinical decision support.
In summary, the UPenn team’s CAMI tool demonstrates promising accuracy in identifying autism in school‑aged children using a simple video test. While the system is not yet a standalone diagnostic, it represents a step toward faster, more objective screening methods that could complement existing clinical practices.
The next phase of the project will focus on larger population studies, adult testing, and further refinement of the algorithm’s performance. The research team has not yet announced plans for regulatory approval or commercial deployment, but the findings suggest potential for broader adoption in educational and clinical settings.
As AI continues to permeate healthcare, tools like CAMI illustrate how machine‑learning models can augment traditional assessment methods, potentially improving access to early diagnosis and intervention for children with autism.
