Chatbots such as ChatGPT may be able to detect early Alzheimer’s by analyzing language patterns, according to groundbreaking research from Drexel University.

The study, published in the digital journal PLOS Digital Health, explained that through language analysis of input text, ChatGPT was able to detect early Alzheimer’s disease (AD) with 80% accuracy.

“We know from ongoing research that the cognitive effects of Alzheimer’s Disease can manifest themselves in language production,” explained Dr. Hualou Liang, the study’s co-author and a professor at Drexel University School of Biomedical Engineering, Science and Health Systems.

Since 60-80% of dementia patients develop language impairment, early Alzheimer’s diagnosing methods involve “acoustic analysis.” Programs detect “subtle cues” within a person’s speech, such as pauses, forgetting the meaning of words, and pronunciation mistakes.

These acoustic programs have a baseline accuracy of 64%, the study found. This means that ChatGPT may outperform conventional methods by 16 percentage points.

Through machine learning, ChatGPT analyzed sentence structure and created a profile for Alzheimer’s speech called an “embedding.” Then, researchers took the embedding and retrained ChatGPT to create GPT-3, a more sophisticated version of ChatGPT specifically designed to detect Alzheimer’s.

“These results, all together, suggest that GPT-3-based text embedding is a promising approach for AD assessment and has the potential to improve early diagnosis of dementia,” the study continued.

Although there is no cure for Alzheimer’s disease, early diagnosis gives a patient more options for support and therapy.

“Our proof-of-concept shows that this could be a simple, accessible, and adequately sensitive tool for community-based testing,” Liang said. “This could be very useful for early screening and risk assessment before a clinical diagnosis.”

Beyond chatbots’ potential to diagnose early Alzheimer’s, a new blood test developed by the University of Pittsburgh is also a promising step toward early Alzheimer’s detection, as reported by The Dallas Express. Unlike a blood test, however, chatbot assessments may be completed remotely at low cost.

Over 400,000 Texans have been diagnosed with Alzheimer’s. Although there is no cure for this disease, obesity and diabetes may be linked to the development of dementia. Therefore, weight management may help prevent it.

As the population ages and obesity rates increase, Texas Health and Human Services expects Alzheimer’s patients to double by 2050.