Natural Language Research Articles

Perspectives on Using Artificial Intelligence to Derive Social Determinants of Health Data From Medical Records in Canada: Large Multijurisdictional Qualitative Study.

Data on the social determinants of health could be used to improve care, support quality improvement initiatives, and track progress toward health equity. However, this data collection is not widespread. Artificial intelligence (AI), specifically natural language processing and machine learning, could be used to derive social determinants of health data from electronic medical records. This could reduce the time and resources required to obtain social determinants of health data. This study aimed to understand perspectives of a diverse sample of Canadians on the use of AI to derive social determinants of health information from electronic medical record data, including benefits and concerns. Using a qualitative description approach, in-depth interviews were conducted with 195 participants purposefully recruited from Ontario, Newfoundland and Labrador, Manitoba, and Saskatchewan. Transcripts were analyzed using an inductive and deductive content analysis. A total of 4 themes were identified. First, AI was described as the inevitable future, facilitating more efficient, accessible social determinants of health information and use in primary care. Second, participants expressed concerns about potential health care harms and a distrust in AI and public systems. Third, some participants indicated that AI could lead to a loss of the human touch in health care, emphasizing a preference for strong relationships with providers and individualized care. Fourth, participants described the critical importance of consent and the need for strong safeguards to protect patient data and trust. These findings provide important considerations for the use of AI in health care, and particularly when health care administrators and decision makers seek to derive social determinants of health data.

Unpacking Sarcasm: A Contextual and Transformer-Based Approach for Improved Detection

Sarcasm detection is a crucial task in natural language processing (NLP), particularly in sentiment analysis and opinion mining, where sarcasm can distort sentiment interpretation. Accurately identifying sarcasm remains challenging due to its context-dependent nature and linguistic complexity across informal text sources like social media and conversational dialogues. This study utilizes three benchmark datasets, namely, News Headlines, Mustard, and Reddit (SARC), which contain diverse sarcastic expressions from headlines, scripted dialogues, and online conversations. The proposed methodology leverages transformer-based models (RoBERTa and DistilBERT), integrating context summarization, metadata extraction, and conversational structure preservation to enhance sarcasm detection. The novelty of this research lies in combining contextual summarization with metadata-enhanced embeddings to improve model interpretability and efficiency. Performance evaluation is based on accuracy, F1 score, and the Jaccard coefficient, ensuring a comprehensive assessment. Experimental results demonstrate that RoBERTa achieves 98.5% accuracy with metadata, while DistilBERT offers a 1.74x speedup, highlighting the trade-off between accuracy and computational efficiency for real-world sarcasm detection applications.

Reward Design Using Large Language Models for Natural Language Explanation of Reinforcement Learning Agent Actions

Reinforcement learning (RL) has found applications across diverse domains; however, it grapples with challenges when formulating reward functions and exhibits low exploration efficiency. Recent studies leveraging large language models (LLMs) have made strides in addressing these issues. However, for RL agents to be practically deployable, elucidating their decision‐making process is crucial for enhancing explainability. We introduce a novel RL approach aimed at alleviating the burden of designing reward functions and facilitating natural language explanations for actions grounded in the agent's decisions. Our method employs two types of agents: a low‐level agent responsible for concrete action selection and a high‐level agent tasked with setting abstract action goals. The high‐level agent undergoes training using a hybrid reward function framework, which incentivizes its actions by comparing them with those generated by an LLM across discretized states. Meanwhile, the training of the low‐level agent is guided by a reward function designed using the EUREKA algorithm. We applied the proposed method to the cart‐pole problem and demonstrated its ability to achieve a learning convergence rate while reducing human effort. Moreover, our approach yields coherent natural language explanations elucidating the rationale behind the agent's actions. © 2025 The Author(s). IEEJ Transactions on Electrical and Electronic Engineering published by Institute of Electrical Engineers of Japan and Wiley Periodicals LLC.

On the State of NLP Approaches to Modeling Depression in Social Media: A Post-COVID-19 Outlook.

Computational approaches to predicting mental health conditions in social media have been substantially explored in the past years. Multiple reviews have been published on this topic, providing the community with comprehensive accounts of the research in this area. Among all mental health conditions, depression is the most widely studied due to its worldwide prevalence. The COVID-19 global pandemic, starting in early 2020, has had a great impact on mental health worldwide. Harsh measures employed by governments to slow the spread of the virus (e.g., lockdowns) and the subsequent economic downturn experienced in many countries have significantly impacted people's lives and mental health. Studies have shown a substantial increase of above 50% in the rate of depression in the population. In this context, we present a review on natural language processing (NLP) approaches to modeling depression in social media, providing the reader with a post-COVID-19 outlook. This review contributes to the understanding of the impacts of the pandemic on modeling depression in social media. We outline how state-of-the-art approaches and new datasets have been used in the context of the COVID-19 pandemic. Finally, we also discuss ethical issues in collecting and processing mental health data, considering fairness, accountability, and ethics.

BERT-based two-channel neural network model text emotion analysis

Introduction: Text emotion analysis, or sentiment analysis, is a crucial area in natural language processing (NLP) focused on identifying emotions within textual data. Challenges in this field include disappearing gradients, information loss, and the lack of contextual semantics.Methods: To address these challenges, we propose a BERT-based model utilizing a two-channel neural network for enhanced emotion classification. The model transforms text into word vectors using BERT, which excels in capturing contextual information. The architecture includes Augmented Recurrent Neural Networks-mutated Unidirectional Long Short-Term Memory (ARNN-Uni-LSTM) to extract local semantic features and capture long-range dependencies. Preprocessing involved tokenization and Word2Vec on publicly available text emotion datasets. The first channel employs ARNN for local feature extraction, while the second uses Uni-LSTM for broader context.Results: Experiments conducted in Python demonstrated that our model outperformed traditional methods, achieving precision of 97.18%, recall of 94.56%, and an F1 score of 96.26%.Conclusions: The BERT-based model shows significant promise for applications such as customer feedback analysis, social media monitoring, and mental health diagnostics, offering a foundation for advanced emotion recognition systems.

A dual adversarial structure of generative adversarial network for nature language generation

PurposeRecently, due to the practicability in several domains, generative adversarial network (GAN) has successfully been adopted in the field of natural language generation (NLG). The purpose of this paper focuses on improving the quality of text and generating sequences similar to human writing for several real applications.Design/methodology/approachA novel model, GAN2, is developed based on a GAN with dual adversarial architecture. We train the generator by an internal discriminator with a beam search technique to improve the quality of generated sequences. Then, we enhance the generator with an external discriminator to optimize and strengthen the learning process of sequence generation.FindingsThe proposed GAN2 model could be utilized in widespread applications, such as chatbots, machine translation and image description. By the proposed dual adversarial structure, we significantly improve the quality of the generated text. The average and top-1 metrics, such as NLL, BLEU and ROUGE, are used to measure the generated sentences from the GAN2 model over all baselines. Several experiments are conducted to demonstrate the performance and superiority of the proposed model compared with the state-of-the-art methods on numerous evaluation metrics.Originality/valueGenerally, reward sparsity and mode collapse are two main challenging issues when adopt GAN to real NLG applications. In this study, GAN2 exploits a dual adversarial architecture which facilitates the learning process in the early training stage for solving the problem of reward sparsity. The occurrence of mode collapse also could be reduced in the later training stage with the introduced comparative discriminator by avoiding high rewards for training in a specific mode. Furthermore, the proposed model is applied to several synthetic and real datasets to show the practicability and exhibit great generalization with all discussed metrics.

Compliance Evaluation with ChatGPT for Diagnosis and Treatment in Patients Brought to the ED with a Preliminary Diagnosis of Stroke

Compliance Evaluation with ChatGPT for Diagnosis and Treatment in Patients Brought to the ED with a Preliminary Diagnosis of Stroke

Enterprise Architecture in the Age of Generative AI: Adapting ERP Systems for Next-Generation Automation

Enterprise architecture is experiencing a profound transformation through the integration of generative artificial intelligence into Enterprise Resource Planning (ERP) systems, fundamentally reshaping how organizations approach automation, decision-making, and strategic planning. This article examines the architectural implications of incorporating AI capabilities into ERP frameworks, focusing on three key dimensions: predictive analytics for enhanced forecasting and risk management, intelligent process automation for operational efficiency, and strategic decision support through natural language processing. Drawing from industry implementations and architectural patterns, this article explores the challenges and opportunities in designing resilient AI-enabled ERP systems that balance innovation with enterprise constraints. The discussion encompasses critical considerations for enterprise architects, including data privacy, integration complexity, and governance frameworks, while providing actionable insights for organizations transitioning to next-generation ERP architectures. This article suggests that successful AI integration in ERP systems requires a holistic architectural approach that aligns technological capabilities with organizational objectives, supported by robust governance mechanisms and clear implementation strategies. This article contributes to the growing body of knowledge on enterprise architecture evolution in the context of emerging AI technologies, offering practical guidance for architects and decision-makers navigating this transformative landscape.

Using Natural Language Processing Methods to Build the Hypersexuality in Bipolar Reddit Corpus: Infodemiology Study of Reddit.

Bipolar is a severe mental health condition affecting at least 2% of the global population, with clinical observations suggesting that individuals experiencing elevated mood states, such as mania or hypomania, may have an increased propensity for engaging in risk-taking behaviors, including hypersexuality. Hypersexuality has historically been stigmatized in society and in health care provision, which makes it more difficult for service users to talk about their behaviors. There is a need for greater understanding of hypersexuality to develop better evidence-based treatment, support, and training for health professionals. This study aimed to develop and assess effective methodologies for identifying posts on Reddit related to hypersexuality posted by people with a self-reported bipolar diagnosis. Using natural language processing techniques, this research presents a specialized dataset, the Talking About Bipolar on Reddit Corpus (TABoRC). We used various computational tools to filter and categorize posts that mentioned hypersexuality, forming the Hypersexuality in Bipolar Reddit Corpus (HiB-RC). This paper introduces a novel methodology for detecting hypersexuality-related conversations on Reddit and offers both methodological insights and preliminary findings, laying the groundwork for further research in this emerging field. A toolbox of computational linguistic methods was used to create the corpora and infer demographic variables for the Redditors in the dataset. The key psychological domains in the corpus were measured using Linguistic Inquiry and Word Count, and a topic model was built using BERTopic to identify salient language clusters. This paper also discusses ethical considerations associated with this type of analysis. The TABoRC is a corpus of 6,679,485 posts from 5177 Redditors, and the HiB-RC is a corpus totaling 2146 posts from 816 Redditors. The results demonstrate that, between 2012 and 2021, there was a 91.65% average yearly increase in posts in the HiB-RC (SD 119.6%) compared to 48.14% in the TABoRC (SD 51.2%) and an 86.97% average yearly increase in users (SD 93.8%) compared to 27.17% in the TABoRC (SD 38.7%). These statistics suggest that there was an increase in posting activity related to hypersexuality that exceeded the increase in general Reddit use over the same period. Several key psychological domains were identified as significant in the HiB-RC (P<.001), including more negative tone, more discussion of sex, and less discussion of wellness compared to the TABoRC. Finally, BERTopic was used to identify 9 key topics from the dataset. Hypersexuality is an important symptom that is discussed by people with bipolar on Reddit and needs to be systematically recognized as a symptom of this illness. This research demonstrates the utility of a computational linguistic framework and offers a high-level overview of hypersexuality in bipolar, providing empirical evidence that paves the way for a deeper understanding of hypersexuality from a lived experience perspective.

IrAE-GPT: Leveraging large language models to identify immune-related adverse events in electronic health records and clinical trial datasets.

Large language models (LLMs) have emerged as transformative technologies, revolutionizing natural language understanding and generation across various domains, including medicine. In this study, we investigated the capabilities, limitations, and generalizability of Generative Pre-trained Transformer (GPT) models in analyzing unstructured patient notes from large healthcare datasets to identify immune-related adverse events (irAEs) associated with the use of immune checkpoint inhibitor (ICI) therapy. We evaluated the performance of GPT-3.5, GPT-4, and GPT-4o models on manually annotated datasets of patients receiving ICI therapy, sampled from two electronic health record (EHR) systems and seven clinical trials. A zero-shot prompt was designed to exhaustively identify irAEs at the patient level (main analysis) and the note level (secondary analysis). The LLM-based system followed a multi-label classification approach to identify any combination of irAEs associated with individual patients or clinical notes. System evaluation was conducted for each available irAE as well as for broader categories of irAEs classified at the organ level. Our analysis included 442 patients across three institutions. The most common irAEs manually identified in the patient datasets included pneumonitis (N=64), colitis (N=56), rash (N=32), and hepatitis (N=28). Overall, GPT models achieved high sensitivity and specificity but only moderate positive predictive values, reflecting a potential bias towards overpredicting irAE outcomes. GPT-4o achieved the highest F1 and micro-averaged F1 scores for both patient-level and note-level evaluations. Highest performance was observed in the hematological (F1 range=1.0-1.0), gastrointestinal (F1 range=0.81-0.85), and musculoskeletal and rheumatologic (F1 range=0.67-1.0) irAE categories. Error analysis uncovered substantial limitations of GPT models in handling textual causation, where adverse events should not only be accurately identified in clinical text but also causally linked to immune checkpoint inhibitors. The GPT models demonstrated generalizable abilities in identifying irAEs across EHRs and clinical trial reports. Using GPT models to automate adverse event detection in large healthcare datasets will reduce the burden on physicians and healthcare professionals by eliminating the need for manual review. This will strengthen safety monitoring and lead to improved patient care.

Graph Foundation Models: Concepts, Opportunities and Challenges.

Foundation models have emerged as critical components in a variety of artificial intelligence applications, and showcase significant success in natural language processing and several other domains. Meanwhile, the field of graph machine learning is witnessing a paradigm transition from shallow methods to more sophisticated deep learning approaches. The capabilities of foundation models in generalization and adaptation motivate graph machine learning researchers to discuss the potential of developing a new graph learning paradigm. This paradigm envisions models that are pre-trained on extensive graph data and can be adapted for various graph tasks. Despite this burgeoning interest, there is a noticeable lack of clear definitions and systematic analyses pertaining to this new domain. To this end, this article introduces the concept of Graph Foundation Models (GFMs), and offers an exhaustive explanation of their key characteristics and underlying technologies. We proceed to classify the existing work related to GFMs into three distinct categories, based on their dependence on graph neural networks and large language models. In addition to providing a thorough review of the current state of GFMs, this article also outlooks potential avenues for future research in this rapidly evolving domain.

A machine learning language model approach to evaluating mental health awareness content across Spanish- and English-language social media posts on Twitter.

Mental health information appears on social media in varying levels of quality and may or may not be productive information to users, particularly in relation to healthcare decision-making and community living among diverse populations coping with mental health problems. To better understand the mental health landscape on social media, this study validated a language model approach to evaluating the availability and sentiment of mental health awareness content across Spanish- and English-language social media posts on Twitter (currently X) to inform future mental health communication guidelines. A comprehensive list of mental health awareness hashtags in Spanish and English was developed by bilingual investigators to download tweets containing these hashtags in both languages from the Twitter Academic API from 09/19/22 - 10/10/22. Data extraction and cleaning of duplicate tweets resulted in a final sample of 28,268 Spanish and 205,774 English tweets for sentiment and structural topic analysis across the two languages. Fifteen unique topics emerged for both Spanish and English tweets including overlapping themes of awareness, self-care, lived experience, and service providers. Topics in Spanish tweets were more often significantly associated with negative emotions compared to English tweets. Yet English tweets also included misappropriation of mental health labels to make political statements and market products. Mental health awareness content on Twitter appears not to be consistently available or aligned with clinical values, disadvantaging Spanish-language social media users, possibly leading to divergent priorities concerning population mental health. Nevertheless, natural language processing techniques offers a viable method to further understand unequal mental health awareness content across various language and cultural social media.

AI-Powered Education: Transforming Learning through Personalized and Scalable Solutions

The rapid evolution of Artificial Intelligence (AI) has profoundly influenced various sectors, with education emerging as a pivotal area of transformation. The integration of AI into educational systems is redefining teaching methodologies, learning experiences, and administrative efficiencies. However, this intersection of AI and education faces significant challenges, including disparities in access, ethical concerns, and the lack of standardized frameworks for implementation. To address these challenges, this paper proposes a comprehensive AI-powered educational framework designed to personalize learning experiences and scale educational delivery efficiently. The framework incorporates a multi-layered architecture consisting of intelligent tutoring systems, adaptive learning platforms, and automated assessment tools. These components are designed to leverage AI algorithms such as natural language processing, predictive analytics, and machine learning to analyze student data, identify learning gaps, and deliver customized content. The proposed solution was evaluated through case studies and pilot implementations, demonstrating improved learner engagement, enhanced knowledge retention, and optimized resource utilization. Key findings include a 25% improvement in learning outcomes in personalized environments and increased teacher productivity by automating repetitive tasks. This research contributes to the field by offering a scalable and practical model for integrating AI into educational systems. It highlights ethical considerations, emphasizes the importance of inclusivity, and underscores the need for interdisciplinary collaboration. Finally, the paper presents actionable recommendations, including policy guidelines, strategies for addressing equity challenges, and a roadmap for future research. These recommendations aim to guide educators, technologists, and policymakers in harnessing the full potential of AI to create more equitable and effective learning ecosystems.

Advanced NLP-driven predictive modeling for tailored treatment strategies in gastrointestinal cancer.

Gastrointestinal cancer represents a significant health burden, necessitating innovative approaches for personalized treatment. This study aims to develop an advanced natural language processing (NLP)-driven predictive modeling framework for tailored treatment strategies in gastrointestinal cancer, leveraging the capabilities of deep learning. The Resilient Adam Algorithm-driven Versatile Long-Short Term Memory (RAA-VLSTM) model is proposed to analyze comprehensive clinical data. The dataset comprises extensive electronic health records (EHRs) from multiple healthcare centers, focusing on patient demographics, clinical history, treatment outcomes, and genetic factors. Data preprocessing employs techniques such as tokenization, normalization, and stop-word removal to ensure effective representation of textual data. For feature extraction, state-of-the-art word embeddings are utilized to enhance model performance. The proposed framework outlines a comprehensive process: data collection from EHRs, preprocessing to prepare the data for analysis, and employing NLP techniques to extract meaningful features. The RAA optimization algorithm significantly improves training efficiency by adapting learning rates for each parameter, addressing common issues in gradient descent. This optimization enhances feature learning from sequential clinical data, enabling accurate predictions of treatment responses and outcomes. The overall performance in terms of F1-score (89.4%), accuracy (92.5%), recall (88.7%), and precision (90.1%). Preliminary results demonstrate the model's strong predictive capabilities, achieving high accuracy in predicting treatment outcomes, thereby suggesting its potential to improve individualized care. In conclusion, this study establishes a robust foundation for employing advanced NLP and machine learning techniques in the management of gastrointestinal cancer, paving the way for future research and clinical applications.

Natural Language Understanding and Inference with MLLM in Visual Question Answering: A Survey

Visual Question Answering (VQA) is a challenge task that combines natural language processing and computer vision techniques and gradually becomes a benchmark test task in multimodal large language models (MLLMs). The goal of our survey is to provide an overview of the development of VQA and a detailed description of the latest models with high timeliness. This survey gives an up-to-date synthesis of natural language understanding of images and text, as well as the knowledge reasoning module based on image-question information on the core VQA tasks. In addition, we elaborate on recent advances in extracting and fusing modal information with vision-language pretraining models and multimodal large language models in VQA. We also exhaustively review the progress of knowledge reasoning in VQA by detailing the extraction of internal knowledge and the introduction of external knowledge. Finally, we present the datasets of VQA and different evaluation metrics and discuss possible directions for future work.

The Impact of the Human Factor on Communication During a Collision Situation in Maritime Navigation

In this paper, the authors draw attention to the significant impact of the human factor during collision situations in maritime navigation. The problems in the communication process between navigators are so excessive that the authors propose automatic communication. This is an alternative method to the current one. The presented system comprehensively performs communication tasks during a sea voyage. To reach the mentioned goal, AI methods of natural language processing and additional properties of metaontology (ontology supplemented with objective functions) are applied. Dedicated to maritime transport applications, the model for translating a natural language into an ontology consists of multiple steps and uses AI methods of classification for the recognition of a message from the ship’s bridge. The reverse model is also multi-stage and uses a created rule-based knowledge base to create natural-language sentences built on the basis of the ontology. Validation of the model’s accuracy results was conducted through accuracy assessment coefficients for information classification, commonly used in science. Receiver operating characteristic (ROC) curves represent the results in the datasets. The presented solution of the designed architecture of the system as well as algorithms developed in the software prototype confirmed the correctness of the assumptions in the described study. The authors demonstrated that it is feasible to successfully apply metaontology and machine learning methods in the proposed prototype software for ship-to-ship communication.

Semantic Role Labelling: A Systematic Review of Approaches, Challenges, and Trends for English and Indian Languages

ABSTRACTThis systematic review looks at the advances, trends, and challenges within semantic role labelling (SRL) for both English and Indian languages. SRL stands as a pivotal undertaking in the realm of natural language processing (NLP), entailing the identification of semantic connections between predicates and their corresponding arguments in a given sentence. The synthesis of findings from publicly available NLP repositories in this review sheds light on the progression of SRL methodologies and their use across various linguistic contexts. The investigation examines the distinct hurdles presented by Indian languages, which are characterised by their morphological complexity and syntactic variability, juxtaposed with the more widely studied English language. Furthermore, we perform an analysis of the impact of sophisticated machine learning algorithms, particularly deep learning, on enhancing SRL efficacy across these languages. The review identifies key research gaps and proposes future research pathways to address the complex nature of SRL in multilingual environments. By offering a comprehensive overview of the evolutionary trajectory of SRL research, the primary objective of this article is to contribute to the advancement of more resilient and adaptable NLP systems capable of accommodating a myriad of languages.

A systematic review of natural language processing applications in Trauma & Orthopaedics.

Prevalence of artificial intelligence (AI) algorithms within the Trauma & Orthopaedics (T&O) literature has greatly increased over the last ten years. One increasingly explored aspect of AI is the automated interpretation of free-text data often prevalent in electronic medical records (known as natural language processing (NLP)). We set out to review the current evidence for applications of NLP methodology in T&O, including assessment of study design and reporting. MEDLINE, Allied and Complementary Medicine (AMED), Excerpta Medica Database (EMBASE), and Cochrane Central Register of Controlled Trials (CENTRAL) were screened for studies pertaining to NLP in T&O from database inception to 31 December 2023. An additional grey literature search was performed. NLP quality assessment followed the criteria outlined by Farrow et al in 2021 with two independent reviewers (classification as absent, incomplete, or complete). Reporting was performed according to the Synthesis-Without Meta-Analysis (SWiM) guidelines. The review protocol was registered on the Prospective Register of Systematic Reviews (PROSPERO; registration no. CRD42022291714). The final review included 31 articles (published between 2012 and 2021). The most common subspeciality areas included trauma, arthroplasty, and spine; 13% (4/31) related to online reviews/social media, 42% (13/31) to clinical notes/operation notes, 42% (13/31) to radiology reports, and 3% (1/31) to systematic review. According to the reporting criteria, 16% (5/31) were considered good quality, 74% (23/31) average quality, and 6% (2/31) poor quality. The most commonly absent reporting criteria were evaluation of missing data (26/31), sample size calculation (31/31), and external validation of the study results (29/31 papers). Code and data availability were also poorly documented in most studies. Application of NLP is becoming increasingly common in T&O; however, published article quality is mixed, with few high-quality studies. There are key consistent deficiencies in published work relating to NLP which ultimately influence the potential for clinical application. Open science is an important part of research transparency that should be encouraged in NLP algorithm development and reporting.

Improving Learning and Teaching Methods Using Chatgpt in Higher Education: A Case Study of IBB University

ChatGPT’s impressive natural language generation capabilities represent a major advance in practical applications of AI-generated information. The potential use cases for ChatGPT are wide-ranging, including assessment setup, translation tasks, and custom source code creation. It can also handle more complex aspects of scientific writing, such as summarizing the literature and paraphrasing text. The difficulties and challenges of using ChatGPT in learning and teaching methods vary from one society to another and from one country to another, and there are many factors. Therefore, this paper aims to investigate the possibility of improving learning and teaching methods using ChatGPT in higher education. It presents the main challenges and notable opportunities that arise from the introduction of ChatGPT in the context of higher education. In line with the aim, the question guiding the study is: “How can learning and teaching methods be improved using ChatGPT in higher education?” To answer this question, a qualitative exploratory case study was conducted at the Department of Information Technology, Faculty of Science, IBB University. An electronic questionnaire containing closed questions was sent to all students. 93 responses were obtained via email, which constituted the sample of participants in this study. The primary contribution of this study is to provide a proposal on how to effectively integrate ChatGPT into higher education and improve learning and teaching methods in key areas.

Natural Language Processing Technologies for Public Health in Africa: Scoping Review.

Natural language processing (NLP) has the potential to promote public health. However, applying these technologies in African health systems faces challenges, including limited digital and computational resources to support the continent's diverse languages and needs. This scoping review maps the evidence on NLP technologies for public health in Africa, addressing the following research questions: (1) What public health needs are being addressed by NLP technologies in Africa, and what unmet needs remain? (2) What factors influence the availability of public health NLP technologies across African countries and languages? (3) What stages of deployment have these technologies reached, and to what extent have they been integrated into health systems? (4) What measurable impact has these technologies had on public health outcomes, where such data are available? (5) What recommendations have been proposed to enhance the quality, cost, and accessibility of health-related NLP technologies in Africa? This scoping review includes academic studies published between January 1, 2013, and October 3, 2024. A systematic search was conducted across databases, including MEDLINE via PubMed, ACL Anthology, Scopus, IEEE Xplore, and ACM Digital Library, supplemented by gray literature searches. Data were extracted and the NLP technology functions were mapped to the World Health Organization's list of essential public health functions and the United Nations' sustainable development goals (SDGs). The extracted data were analyzed to identify trends, gaps, and areas for future research. This scoping review follows the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews) reporting guidelines, and its protocol is publicly available. Of 2186 citations screened, 54 studies were included. While existing NLP technologies support a subset of essential public health functions and SDGs, language coverage remains uneven, with limited support for widely spoken African languages, such as Kiswahili, Yoruba, Igbo, and Zulu, and no support for most of Africa's >2000 languages. Most technologies are in prototyping phases, with only one fully deployed chatbot addressing vaccine hesitancy. Evidence of measurable impact is limited, with 15% (8/54) studies attempting health-related evaluations and 4% (2/54) demonstrating positive public health outcomes, including improved participants' mood and increased vaccine intentions. Recommendations include expanding language coverage, targeting local health needs, enhancing trust, integrating solutions into health systems, and adopting participatory design approaches. The gray literature reveals industry- and nongovernmental organizations-led projects focused on deployable NLP applications. However, these projects tend to support only a few major languages and specific use cases, indicating a narrower scope than academic research. Despite growth in NLP research for public health, major gaps remain in deployment, linguistic inclusivity, and health outcome evaluation. Future research should prioritize cross-sectoral and needs-based approaches that engage local communities, align with African health systems, and incorporate rigorous evaluations to enhance public health outcomes. RR2-doi:10.1101/2024.07.02.24309815.