English Queries Research Articles

Cross-Linguistic Evaluation of Generative AI Models for Diabetes and Endocrine Queries

Background and Aims: Endocrine and metabolic disorders, including diabetes mellitus (DM), pose major global health challenges. Generative artificial intelligence (genAI) models are increasingly used for patient self-help. This study aimed to evaluate the performance of two genAI models, ChatGPT and Microsoft Copilot, in addressing endocrine-related queries in English and Arabic. Materials and Methods: This descriptive study adhered to the METRICS checklist for genAI-based healthcare studies, comparing responses from ChatGPT-4o and Microsoft Copilot to 20 endocrine-related queries in English and Arabic (15 DM queries in addition to five endocrine queries). The responses were evaluated using the CLEAR tool, which assessed completeness, accuracy, and relevance/appropriateness. Three endocrinology experts independently evaluated the genAI outputs. Results: Per language and model, a total of 80 responses were assessed. Inter-rater reliability was high with Intraclass Correlation Coefficient=0.832. ChatGPT-4o consistently outperformed Microsoft Copilot, earning 'Excellent' ratings in English and ‘Very good’ in Arabic, while Microsoft Copilot achieved ‘Very good’ ratings in English and ‘Good’ to ‘Very good’ ratings in Arabic. ChatGPT-4o surpassed Microsoft Copilot in completeness (4.38 vs. 3.36, p<.001), accuracy (4.18 vs. 3.83, p=.014), and relevance (4.44 vs. 3.82, p<.001). Performance varied significantly between English and Arabic responses, with p<.001 for completeness, p=.001 for accuracy, p=.012 for relevance, and p<.001 for the overall CLEAR score. No statistically significant differences were found based on the query topic. Conclusions: ChatGPT-4o outperformed Microsoft Copilot in all CLEAR components, but notable language-based disparities were evident. Addressing these limitations is crucial to ensure equitable access to endocrine care for non-English-speaking patients.

Exploring the role of artificial intelligence, large language models: Comparing patient-focused information and clinical decision support capabilities to the gynecologic oncology guidelines.

Gynecologic cancer requires personalized care to improve outcomes. Large language models (LLMs) hold the potential to provide intelligent question-answering with reliable information about medical queries in clear and plain English, which can be understood by both healthcare providers and patients. We aimed to evaluate two freely available LLMs (ChatGPT and Google's Bard) in answering questions regarding the management of gynecologic cancer. The LLMs' performances were evaluated by developing a set questions that addressed common gynecologic oncologic findings from a patient's perspective and more complex questions to elicit recommendations from a clinician's perspective. Each question was presented to the LLM interface, and the responses generated by the artificial intelligence (AI) model were recorded. The responses were assessed based on the adherence to the National Comprehensive Cancer Network and European Society of Gynecological Oncology guidelines. This evaluation aimed to determine the accuracy and appropriateness of the information provided by LLMs. We showed that the models provided largely appropriate responses to questions regarding common cervical cancer screening tests and BRCA-related questions. Less useful answers were received to complex and controversial gynecologic oncology cases, as assessed by reviewing the common guidelines. ChatGPT and Bard lacked knowledge of regional guideline variations, However, it provided practical and multifaceted advice to patients and caregivers regarding the next steps of management and follow up. We conclude that LLMs may have a role as an adjunct informational tool to improve outcomes.

Language discrepancies in the performance of generative artificial intelligence models: an examination of infectious disease queries in English and Arabic

BackgroundAssessment of artificial intelligence (AI)-based models across languages is crucial to ensure equitable access and accuracy of information in multilingual contexts. This study aimed to compare AI model efficiency in English and Arabic for infectious disease queries.MethodsThe study employed the METRICS checklist for the design and reporting of AI-based studies in healthcare. The AI models tested included ChatGPT-3.5, ChatGPT-4, Bing, and Bard. The queries comprised 15 questions on HIV/AIDS, tuberculosis, malaria, COVID-19, and influenza. The AI-generated content was assessed by two bilingual experts using the validated CLEAR tool.ResultsIn comparing AI models’ performance in English and Arabic for infectious disease queries, variability was noted. English queries showed consistently superior performance, with Bard leading, followed by Bing, ChatGPT-4, and ChatGPT-3.5 (P = .012). The same trend was observed in Arabic, albeit without statistical significance (P = .082). Stratified analysis revealed higher scores for English in most CLEAR components, notably in completeness, accuracy, appropriateness, and relevance, especially with ChatGPT-3.5 and Bard. Across the five infectious disease topics, English outperformed Arabic, except for flu queries in Bing and Bard. The four AI models’ performance in English was rated as “excellent”, significantly outperforming their “above-average” Arabic counterparts (P = .002).ConclusionsDisparity in AI model performance was noticed between English and Arabic in response to infectious disease queries. This language variation can negatively impact the quality of health content delivered by AI models among native speakers of Arabic. This issue is recommended to be addressed by AI developers, with the ultimate goal of enhancing health outcomes.

Performance of ChatGPT in Answering Clinical Questions on the Practical Guideline of Blepharoptosis.

ChatGPT is a free artificial intelligence (AI) language model developed and released by OpenAI in late 2022. This study aimed to evaluate the performance of ChatGPT to accurately answer clinical questions (CQs) on the Guideline for the Management of Blepharoptosis published by the American Society of Plastic Surgeons (ASPS) in 2022. CQs in the guideline were used as question sources in both English and Japanese. For each question, ChatGPT provided answers for CQs, evidence quality, recommendation strength, reference match, and answered word counts. We compared the performance of ChatGPT in each component between English and Japanese queries. A total of 11 questions were included in the final analysis, and ChatGPT answered 61.3% of these correctly. ChatGPT demonstrated a higher accuracy rate in English answers for CQs compared to Japanese answers for CQs (76.4% versus 46.4%; p = 0.004) and word counts (123 words versus 35.9 words; p = 0.004). No statistical differences were noted for evidence quality, recommendation strength, and reference match. A total of 697 references were proposed, but only 216 of them (31.0%) existed. ChatGPT demonstrates potential as an adjunctive tool in the management of blepharoptosis. However, it is crucial to recognize that the existing AI model has distinct limitations, and its primary role should be to complement the expertise of medical professionals. Observational study under respected authorities. This journal requires that authors assign a level of evidence to each article. For a full description of these Evidence-Based Medicine ratings, please refer to the Table of Contents or the online Instructions to Authors www.springer.com/00266 .

BiLSTMs and BPE for English to Telugu CLIR

A crucial component of Cross Lingual Information Retrieval (CLIR) is Neural Machine Translation (NMT). NMT performs a good job of transforming queries in the English language into Indian languages. This study focuses on the translation of English queries into Telugu. For translations, the NMT will make use of a parallel corpus. Due to a lack of resources in the Telugu language, it is exceedingly challenging to provide NMT with sizable parallel corpora. Thus, the NMT will encounter an issue known as Out of Vocabulary (OOV). Long Short-Term Memory (LSTM) with Byte Pair Encoding (BPE), which breaks up rare words into subwords and attempts to translate them to solve the OOV issue. Issues such as Named Entity Recognition (NER) continue to plague it. In sequence-to-sequence models, bidirectional LSTMs can solve certain NER challenges. Systems that need to be trained in both directions to recognize named entities can benefit from the use of Bidirectional LSTMs (BiLSTMs). The translation efficiency of NMT with BiLSTMs is significantly higher than normal LSTMs, as indicated by the accuracy metrics and Bilingual Evaluation Understudy (BLEU) score.

Multilingual Question Answering for Malaysia History with Transformer-based Language Model

In natural language processing (NLP), a Question Answering System (QAS) refers to a system or model that is designed to understand and respond to user queries in natural language. As we navigate through the recent advancements in QAS, it can be observed that there is a paradigm shift of the methods used from traditional machine learning and deep learning approaches towards transformer-based language models. While significant progress has been made, the utilization of these models for historical QAS and the development of QAS for Malay language remain largely unexplored. This research aims to bridge the gaps, focusing on developing a Multilingual QAS for history of Malaysia by utilizing a transformer-based language model. The system development process encompasses various stages, including data collection, knowledge representation, data loading and pre-processing, document indexing and storing, and the establishment of a querying pipeline with the retriever and reader. A dataset with a collection of 100 articles, including web blogs related to the history of Malaysia, has been constructed, serving as the knowledge base for the proposed QAS. A significant aspect of this research is the use of the translated dataset in English instead of the raw dataset in Malay. This decision was made to leverage the effectiveness of well-established retriever and reader models that were trained on English data. Moreover, an evaluation dataset comprising 100 question-answer pairs has been created to evaluate the performance of the models. A comparative analysis of six different transformer-based language models, namely DeBERTaV3, BERT, ALBERT, ELECTRA, MiniLM, and RoBERTa, has been conducted, where the effectiveness of the models was examined through a series of experiments to determine the best reader model for the proposed QAS. The experimental results reveal that the proposed QAS achieved the best performance when employing RoBERTa as the reader model. Finally, the proposed QAS was deployed on Discord and equipped with multilingual support through the incorporation of language detection and translation modules, enabling it to handle queries in both Malay and English. Doi: 10.28991/ESJ-2024-08-02-019 Full Text: PDF

227 Spine-tuned Natural Language Models and Bespoke Regular Expression Classifiers for Automated Spinal Surgery Registry Development

INTRODUCTION: Surgical research demands the development of clinical registries, often through time-intensive manual chart review. Natural language processing (NLP) may accelerate registry development, and an ideal automatic registry (autoregistry) algorithm would be highly accurate while requiring minimal manual data annotation. NLP approaches including bespoke Regular Expression (RegEx) classifiers and Large Language Models (LLM) possess distinct strengths and weaknesses and have not been compared in the setting of autoregistry development. METHODS: We used an institutional data lake to retrieve 31,502 neurosurgical operative notes. A standardized set of spinal procedures was chosen for inclusion in the autoregistry. 200 manually annotated notes were used for training and testing purposes. RegEx classifiers were engineered to retrieve procedural info from unprocessed notes. A family of 110-million parameter BERT models, including LLM pre-trained on clinical text, was fine-tuned for the same tasks. We also tested a open-source 7-billion parameter LLM chatbot, Vicuna, without fine-tuning. RESULTS: The RegEx classifiers were able to identify spinal procedures and associated vertebral levels in nearly 99% of operative notes. Fine-tuned LLM identified common procedures (e.g. spinal fusion and laminectomy) with greater than 95% accuracy but performed poorly for rarer procedures (e.g. XLIF, corpectomy) and vertebral body identification. Qualitative evaluation of the Vicuna chatbot showed potential for the same tasks, following iteratively refined prompting. CONCLUSIONS: The goal of autoregistry development is to minimize time- and labor-intensive manual chart review. We found that fine-tuned LLM could not match the accuracy and efficiency of the RegEx classifier. However, LLM may be well-suited to expand existing clinical databases that provide a robust training set. Further work combining NLP approaches will attempt to develop a pipeline for autoregistry development from natural language (plain English) queries.

Retracted: Construction of a 5G Wireless Semantic Web-Assisted English Digital Learning Resource Query System

Retracted: Construction of a 5G Wireless Semantic Web-Assisted English Digital Learning Resource Query System

Information amount, accuracy, and relevance of generative artificial intelligence platforms’ answers regarding learning objectives of medical arthropodology evaluated in English and Korean queries in December 2023: a descriptive study

This study assessed the performance of 6 generative artificial intelligence (AI) platforms on the learning objectives of medical arthropodology in a parasitology class in Korea. We examined the AI platforms’ performance by querying in Korean and English to determine their information amount, accuracy, and relevance in prompts in both languages. From December 15 to 17, 2023, 6 generative AI platforms—Bard, Bing, Claude, Clova X, GPT-4, and Wrtn—were tested on 7 medical arthropodology learning objectives in English and Korean. Clova X and Wrtn are platforms from Korean companies. Responses were evaluated using specific criteria for the English and Korean queries. Bard had abundant information but was fourth in accuracy and relevance. GPT-4, with high information content, ranked first in accuracy and relevance. Clova X was 4th in amount but 2nd in accuracy and relevance. Bing provided less information, with moderate accuracy and relevance. Wrtn’s answers were short, with average accuracy and relevance. Claude AI had reasonable information, but lower accuracy and relevance. The responses in English were superior in all aspects. Clova X was notably optimized for Korean, leading in relevance. In a study of 6 generative AI platforms applied to medical arthropodology, GPT-4 excelled overall, while Clova X, a Korea-based AI product, achieved 100% relevance in Korean queries, the highest among its peers. Utilizing these AI platforms in classrooms improved the authors’ self-efficacy and interest in the subject, offering a positive experience of interacting with generative AI platforms to question and receive information.

Bidirectional LSTMs with Byte Pair Encoding in NMT for CLIR using English and Telugu Parallel Corpus

The Neural Machine Translation (NMT) is very crucial for Cross-Lingual Information Retrieval (CLIR). NMT is effective in translating English language queries to the Telugu Language. In this paper, we are translating English queries to Telugu. The NMT will utilize a parallel corpus for translations. Telugu is a resource-poor language, it is very difficult to supply large amounts of parallel corpus to NMT. So the NMT will have a problem called Out Of Vocabulary (OOV). To overcome this problem Byte Pair Encoding (BPE) is used along with Long Short Term Memory (LSTM), which segments the rare words into sub-words and tries to translate the rare words. It still faces problems like Named Entity Recognition (NER). Some problems of NER can be solved by utilizing bidirectional LSTMs in sequence-to-sequence models. The bidirectional LSTMs (BiLSTMs) will be helpful in training systems in both directions for recognizing the named entities. The accuracy parameters and a BLEU score show the translation quality of NMT with bidirectional LSTMs has slightly more accuracy than regular LSTMs which is considerable.

Specifics of using text and graphical chatbots with artificial intelligence in English language teaching

Introduction. The current development of mobile applications, including the introduction of chatbots using artificial intelligence, further expands the possibilities of the teaching process, its personalization, and the development of an individual educational trajectory. Using the chatbots ChatGPT and Midjourney will make it possible to practice English language skills at one’s own speed, simulating communication with a native speaker, as well as practice the formulation of clear commands for an algorithmized executor. The aim of this paper is to describe the specifics of using chatbots in the process of studying English by undergraduate students. Materials and methods. The experimental study involved 186 students of the Faculty of Linguistics of A.S. Griboyedov Moscow University (Russian Federation). The experimental group attended elective classes, during which the students debated in English with the ChatGPT neural network and generated images using queries in English in the Midjourney network. Methods of mathematical statistics: Pearson's chi-squared test. Results. The practice of using neural networks in teaching a foreign language has demonstrated that a neural network can act as a simulator of a native speaker, as well as a kind of “command executor” obeying simple instructions. The results confirming these statements and the effectiveness of the experimental work are embodied in the positive dynamics of generating creative statements and texts in a foreign language (χ2 = 9.174; p < 0.05). Conclusion. The data obtained in this experiment make it possible to recommend the use of chatbots with artificial intelligence tools and expand the range of exercises with their application. Using chatbots in this capacity is little elaborated in Russian educational practice, and the scientific novelty of the obtained results lies in suggesting a specific methodology of the option for introducing interaction with automatically learning chatbots in the educational process.

PooRaa‐Agri KG: An agricultural knowledge graph‐based simplified multilingual query system

AbstractThe current work proposes PooRaa‐Agri KG, an agricultural knowledge graph‐based simplified multilingual query system that works in real time to provide concise answers for agriculture‐based queries. The proposed approach accommodates real‐time and low‐resource queries in English and Hindi with a novel multi‐stage solution consisting of data pre‐processing, sentence simplification, triplet extraction, knowledge graph generation, sentence reconstruction, query‐to‐reconstructed sentence matching, and machine translation as its sub‐modules. In this work, a novel combination of rule‐based sentence simplification and triplet extraction is carried out resulting in a triplet similarity score of 86.56% for the extracted triplets. This method is superior to the existing triplet extraction method whose triplet similarity score was found to be 60.65%. Further, the proposed work makes use of heuristic rules to reconstruct sentences which when evaluated by human evaluators for meaningfulness and grammar resulted in a score of 3.09/4 and 2.95/4 respectively. To complete end‐to‐end communication in the proposed system, a similarity‐based query answer system is proposed in this work.

Conversion of NLP to SQL Using Machine Learning Approach

Abstract: Now days data is increasing rapidly. There are so many new database tools and technologies are rowing, therefore we can store large data, but the problem is that the technology or an interface which can process data and display the data as per the user request is not familiarized with many of the people. It means many people don’t have proper knowledge of handling database. Actually, NLP to SQL conversion is part of machine learning. The main idea of NLP to SQL conversion is to generate SQL query from natural language. This technique is useful for accessing data from database without having prior knowledge of SQL. This technique can be used by many common people. In this system input is simple English text and query is generated by using POS tagger in python. In this project we are going to implement a system which will generate SQL query from natural language.

Lessons Learned From Monitoring Spanish-Language Vaccine Misinformation During the COVID-19 Pandemic.

COVID-19 vaccine misinformation is a global threat, and digital and social media support its spread. Addressing Spanish-language vaccine misinformation is critical. In 2021, we began a project to increase vaccine confidence and uptake in the United States by assessing and opposing Spanish-language COVID-19 vaccine misinformation circulating in the United States. Analysts identified trending Spanish-language vaccine misinformation each week, and trained journalists provided communications guidance for addressing the misinformation, which we delivered to community organizations via a weekly newsletter. We identified thematic and geographic trends and highlighted lessons learned to inform future efforts to monitor Spanish-language vaccine misinformation. We collected publicly available Spanish- and English-language COVID-19 vaccine misinformation across various media sources (eg, Twitter, Facebook, news, blogs). Analysts identified top trending vaccine misinformation in the Spanish query and compared it with vaccine misinformation in the English query. Analysts examined misinformation to identify its geographic source and dominant conversation themes. From September 2021 through March 2022, analysts flagged 109 pieces of trending Spanish-language COVID-19 vaccine misinformation. Through this work, we found that Spanish-language vaccine misinformation is easily identifiable. Linguistic networks are not distinct, and vaccine misinformation often circulates across English and Spanish queries. Several websites have outsized influence in promoting Spanish-language vaccine misinformation, suggesting that it may be important to focus on a handful of hyperinfluential accounts and websites. Efforts to address Spanish-language vaccine misinformation must incorporate collaboration with local communities and emphasize community building and empowerment. Ultimately, addressing Spanish-language vaccine misinformation is not an issue of data access and knowledge of how to monitor it; it is an issue of prioritization.

Low-dose computed tomography in COVID-19: systematic review

BACKGROUND: The increased number of computed tomography scans during the COVID-19 pandemic has emphasized the task of decreasing radiation exposure of patients, since it is known to be associated with an elevated risk of cancer development. The ALARA (as low as reasonably achievable) principle, proposed by the International Commission on Radiation Protection, should be adhered to in the operation of radiation diagnostics departments, even during the pandemic.
 AIM: To systematize data on the appropriateness and effectiveness of low-dose computed tomography in the diagnosis of lung lesions in COVID-19.
 MATERIALS AND METHODS: Relevant national and foreign literature in scientific libraries PubMed and eLIBRARY, using English and Russian queries low-dose computed tomography and COVID-19, published between 2020 and 2022 were analyzed. Publications were evaluated after assessing the relevance to the review topic by title and abstract analysis. The references were further analyzed to identify articles omitted during the search that may meet the inclusion criteria.
 RESULTS: Published studies summarized the current data on the imaging of COVID-19 lung lesions and the use of computed tomography scans and identified possible options for reducing the effective dose.
 CONCLUSION: We present techniques to reduce radiation exposure during chest computed tomography and preserve high-quality diagnostic images potentially sufficient for reliable detection of COVID-19 signs. Reducing radiation dose is a valid approach to obtain relevant diagnostic information, preserving opportunities for the introduction of advanced computational analysis technologies in clinical practice.

Spontaneous retroperitoneal hematoma with duodenal obstruction with diagnostic use of endoscopic ultrasound: A case series and literature review

Spontaneous retroperitoneal hematoma is rare and can cause duodenal obstruction. We report four cases of spontaneous retroperitoneal hematoma with duodenal obstruction, wherein endoscopic ultrasound was useful for diagnosis. The patients complained of vomiting with stable vital signs. Computed tomography, esophagogastroduodenoscopy, and endoscopic ultrasound findings were similar in all cases. Contrast-enhanced computed tomography revealed a low-density mass around the 2nd to 3rd part of the duodenum. Esophagogastroduodenoscopy showed an edematous, reddish, but non-neoplastic duodenal mucosa with stenosis of the lumen. Endoscopic ultrasound revealed a low-echoic mass around the duodenum and high-echoic floating matter suggesting debris and anechoic areas that indicated a liquid component. These findings suggested hematomas or abscesses. Although pseudoaneurysm of the pancreaticoduodenal artery was suspected in Case 3, we chose conservative treatment because the aneurysm was small. In Case 4, median arcuate ligament syndrome was suspected on angiography. No aneurysms or arteriovenous malformations were found; thus, endovascular embolization was not performed. The patients were treated conservatively and discharged within 3–5 weeks. English literature queries on spontaneous retroperitoneal hematoma with duodenal obstruction in MEDLINE revealed 21 cases in 18 studies. The clinical features of these patients and the present four cases have been discussed.

Cultural Heritage Topics in Online Queries: A Comparison between English- and Polish-Speaking Internet Users

New communication channels and methods for retrieving information can provide increasingly precise data describing how cultural heritage is perceived, protected, promoted, and shared. Many internet users search for cultural-heritage-related topics using online search engines and databases. The purpose of this study was to identify the context and frequency of cultural heritage topics in Google search engine queries. The primary hypothesis was that keywords used in Google searches related to cultural heritage would be much more diversified in English than in Polish, perhaps because Polish has many fewer internet users than English. The keywords were selected because of their frequent use in the research literature, popular science publications, and central and local government strategy documents. The research employed selected online applications. A total of 3690 keywords were collected, with 1634 (44.3%) in Polish and 2056 (55.7%) in English. The numbers of retrieved keywords were similar for all research tools, but an in-depth analysis identified certain differences in the functions of the tools. The “cultural heritage” keyword in Polish (“dziedzictwo kulturowe”) was found mostly in national, regional, and local contexts. English queries included topics related to cultural heritage protection, conservation, restoration, and management and the financial value of cultural heritage. Some queries focused on world cultural heritage. The research shows that Polish-speaking users and English-speaking ones were interested in similar phrases. Therefore, the thought processes of internet users may be independent of their language.

Construction of artificial intelligence-assisted English learning resource query system.

English has become an important tool for China's opening to the outside world and exchanges with other countries. More and more people have the motivation and requirements to learn English, but under the traditional English learning mode and traditional teaching mode, the cultivation of learners' autonomous learning habits is ignored. This article aims to study the construction of artificial intelligence-assisted English learning resource query system and establish the relevant feedback mechanism of retrieval. This article applies this mechanism to the retrieval of learning resources, so as to provide learners with the learning resources they really need and improve learners' learning efficiency. This article proposes to find the relevant knowledge points by extracting the knowledge points of the retrieval content. It realizes the query expansion based on knowledge and then realizes the expansion of retrieval results. It realizes the mapping of knowledge points on the retrieval content, the query and expansion of knowledge points, and the presentation of learning resources of the knowledge point index. It also uses the relevant feedback mechanism to adjust the retrieval results to meet the retrieval needs of learners. The experimental results show that the number of knowledge points can be increased to 2–4 times by query expansion based on English resources. Thus, the number of learning resources of search results can be increased to 3–10 times, the expansion of search results can be realized, and the overall recall will be greatly improved. In this article, the related methods of artificial intelligence are applied to the construction experiment of the English learning resource query system, which has a certain promotion effect on the construction of the system.

Construction of Computer English Corpus Assisted by Internet of Things Information Perception and Interaction Technology.

Aiming at the situation that multiple sensing devices in the sensing layer sense the same sensing quantity, multiple redundant measurement data will be generated. In this paper, an improved Kalman filtering algorithm is used. The algorithm combines with numerical test theory to remove abnormal error data before filtering, thereby reducing the error of filtering result and improving its reliability. The parallel multiprotocol processing mechanism of the gateway is designed, and the IoT gateway has a hierarchical modular processing design for different protocol signals of the heterogeneous perception layer. The web interaction function with good platform versatility is designed to provide a visual interaction interface for the application layer, which is convenient for the application layer terminal to monitor and manage the perception layer equipment. This paper realizes the retrieval and query function of keywords, wrong sentences, and English character errors. This corpus enriches the less existing Madagascar corpus and establishes a separate English character error query database. This paper describes the overall process of corpus from preliminary preparation to establishment and the related use instructions of corpus, makes a statistical analysis of the current corpus data, and puts forward some suggestions for building a computer English corpus according to the analysis results.

A Global Survey of Android Dual-Use Applications Used in Intimate Partner Surveillance

Intimate partner violence (IPV) is a pervasive societal problem that affects millions of people around the world. IPV perpetrators increasingly weaponize digital technologies like mobile applications (“apps”) to spy on, monitor, and harass victims. Surveillance-capable apps can have legitimate use cases, for example, locating children, and are therefore easily available on various mobile app stores like the Google Play Store. Nevertheless, these applications are easily repurposed by abusers to track their victims. The problem of such dual-use apps in IPV is global. However, current understanding of the ecosystem of such apps is limited to English-language apps, potentially limiting its relevance to non-English speaking IPV survivors across the world. In this paper, we study the prevalence of dualuse applications found in 15 languages and 27 countries. We collected 51,868 unique apps in 2020 from the Google Play Store, using queries such as “track wife’s location.” Through a semi-manual analysis of a subset of these apps, we discovered 854 unique dualuse apps, and estimate that among the apps collected from Google Play, 3,988 are dual-use apps. We found notable differences in app search results, suggested queries, and marketed capabilities of dual-use apps across different languages. For instance, we identified that 18% of dual-use apps do not have an English description, and 28% could not be found using English queries. Google Play (cursorily) blocks certain queries referring explicitly to intimate partner surveillance (IPS) to discourage potential abusers, but the blocking efficacy varies across languages. For example, we found that 80% of explicit IPS queries for English are blocked, but none for Bengali, Chinese, Hindi, Malay, Thai, and Vietnamese. Thus, abusers fluent in those languages can evade such blocking with no effort.