This study, presented by Humayun et al., performed a systematic review and meta-analysis on the application of AI models, Woebot, Tess, Wysa, etc., on alleviating anxiety and depression compared to traditional conversational agents (CA). The databases Elsevier, Scopus, PubMed, and Google Scholar were searched dating original research from January 2000 to July 2024 and highlighting keywords such as “Natural Language Processing” and “Psychiatric Disorder” within the context of peer-reviewed, English-only studies. Of the six articles found, 901 participants took part using Cohen’s d as a measure of the difference between the prior to treatment and post-treatment groups. AI models were suggested to be more effective than conventional CA models at calming anxiety and depression in patients with statistically significant Cohen’s d scores of d = 0.62 and d = 0.74 respectively. One study, presented by Prochaska et al., examined the use of Woebot AI system on anxiety and depression stemming from substance abuse such that approximately 734 participants reported fewer instances of substance use. Humayun et al. considered future studies across a longer time period monitoring participant status and noted the limitation of small sample sizes that did not account for heterogeneity across multiple populations.