On a global scale, suicide is recognized as a major cause of preventable death. This has led to the emergence of different solutions for the identification and prevention of suicide. One solution is the use of artificial intelligence algorithms in the determination of suicidal texts for the further prevention of suicide attempt. In this work, the authors present an AI model that can classify texts into suicidal and nonsuicidal. The proposed model was developed using the Keras framework and trained on data sourced from a publicly available Kaggle repository. The trained model yielded high results, with performance metrics showing an F1 of 0.92 and an AUC close to 0.98. This work provides the design of the model along with data selection and data processing techniques. The paper also explores various approaches in machine and deep learning, natural language processing that are currently used to address mental health issues. In addition, it reviews and critically discusses the contributions and findings of other researchers in this field. In future works, the authors will integrate the developed model into automated real-time social media monitoring systems.