This article presents a systematic review of advances in pre-trained models for real-time translation from Spanish to Colombian Sign Language (CSL). As artificial intelligence and deep learning technologies, such as convolutional neural networks (CNN) and recurrent neural networks (RNN), have evolved, innovative solutions have been developed to overcome the communication barriers faced by deaf people. This paper analyzes the effectiveness of these models, highlighting specific cases such as the CNN-based system that achieved an 88% recognition rate for CSL gestures, and other approaches that use transfer learning techniques to improve translation accuracy and speed. Additionally, it addresses the significant impacts of the lack of effective communication on the education and employment of deaf people, and how automatic translation technologies can enhance social and labor inclusion. Furthermore, it explores the adaptation of models to new users through meta-learning and the implementation of continuous feedback systems to improve accuracy and usability. The findings suggest a promising future for the integration of these technologies into daily life, promoting more inclusive and accessible communication for the deaf community in Colombia.