Cracking the Autism Code: How AI Became a Kid's Best Friend

Researchers are currently investigating the potential of utilizing artificial intelligence (AI) to identify autism and aid individuals with autism in enhancing their social, communicative, and emotional capabilities. The application of AI in diagnosing autism has become a reality, and ongoing developments in AI-based therapies show promise. Several AI-driven applications are now accessible for download by smartphone users. AI can also respond to an individual's unique strengths and challenges with predefined outcomes, much like a teacher or therapist would (1^✔ ✔Trusted Source
Can Autism Be Diagnosed with Artificial Intelligence? A Narrative Review

Go to source).

Advertisement

AI’s Role in Diagnosing and Treating Autism

It's important to note that AI has not yet reached the level depicted in science fiction.

Receiving a diagnosis can take considerable time, potentially leading to delays in accessing necessary therapies and services. AI can aid in early identification of autism traits, particularly in individuals with lower support needs (sometimes referred to as high-functioning), which can be challenging.

Several factors contribute to these delays. Autism does not have a singular defining trait, and certain traits associated with autism can also manifest in unrelated disorders or personality differences. Evaluators might struggle to determine whether a specific behavior is indicative of an autistic pattern or a personal idiosyncrasy. Both evaluators and parents might be reluctant to label a child until they are absolutely certain of the diagnosis.

A form of AI known as "deep learning" might possess a greater ability than humans to recognize relevant patterns. Deep learning, a subset of machine learning based on artificial neural networks, could serve as a valuable resource for confirming a diagnosis or suggesting the need for further evaluation.

Several companies are pioneering methods for diagnosing autism using AI and similar technologies.

Advertisement

Behavior Imaging in Autism

Boise, Idaho-based Behavior Imaging employs a system called the Naturalistic Observation Diagnostic Assessment. This app enables parents to upload videos of their children for observation.

Clinicians analyze these videos remotely to make diagnoses. More recently, the company has begun training AI algorithms to observe and categorize behaviors. While these algorithms would not diagnose children, they could guide clinicians toward specific behaviors that might otherwise be overlooked.

Cognoa, headquartered in Palo Alto, California, offers an AI-assisted screening tool for autism. This mobile app allows parents to conduct assessments without involving a trained evaluator. The app reviews answers to multiple-choice questions and analyzes videos of the child.

However, despite the interest and some utilization of AI to support diagnoses, there is limited support for the notion that AI alone can deliver a dependable autism diagnosis.

Advertisement

Utilizing Robots for Autism Therapy

Individuals with autism frequently encounter challenges in human interactions. Social expectations, sensory difficulties, communication issues, and attention-related problems can impede optimal development. To address these challenges, innovative groups have explored ways to leverage AI in teaching and engaging individuals on the autism spectrum.

One notable approach involves developing and training robots to interact with autistic children. These robots aim to provide practice in recognizing facial expressions, engaging socially, and appropriately responding to social cues.

AI Applications for Autism

Compared to high-end robots, AI-driven applications are more cost-effective and easily integrated into homes, schools, and therapy settings. Numerous autism apps are available, supporting behavioral therapy and learning. However, most of these apps are relatively simple, logic-based tools that follow predefined rules and offer rewards based on compliance.

Distinguishing AI from conventional tech logic, Dipayana explains, "Interaction might commence with a standard response, but then the model evolves." AI apps use exercises to assist users in calming down or responding appropriately. Depending on the child's mood, the model provides exercises and adapts based on the child's responses. This framework allows AI to learn in a manner more reminiscent of human cognition.

One such AI app is the Manatee app, which is accessible as a free iPhone download. Clinical psychologists design the app's goals. It's recommended for parents to engage in the activities with their children. The app follows a step-by-step progression from simple to advanced skills, providing guidance and emphasizing parental involvement.

Limits of AI in Autism Treatment

AI is a nascent tool in the field of autism treatment, and research on its outcomes is currently limited. While AI-driven robots and apps possess the capacity to support learning in children, they have certain limitations.

For instance,the high production and utilization costs associated with robots. App users also need to possess reading skills and the ability to follow instructions.

Apps are designed to teach specific skills, such as appropriate social communication, facial expression recognition, and eye contact. While some children might engage more readily with robots, it remains uncertain whether these skills can be transferred to interactions with human peers.

The integration of apps into typical settings is not yet widespread. Although some therapists and schools are adopting this technology, there is still progress to be made.

Reference:

1. Can Autism Be Diagnosed with Artificial Intelligence? A Narrative Review - (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8618159/)

Source-Medindia