Membership in ACM includes a subscription to Communications of the ACM (CACM), the computing industry's most trusted source for staying connected to the world of advanced computing. The emergence of ChatGPT marked a transformative milestone for artificial intelligence (AI), showcasing the remarkable potential of large language models (LLMs) to generate human-like text. This wave of innovation has revolutionized how we interact with technology, seamlessly integrating LLMs into everyday tasks such as vacation planning, email drafting, and content creation. While English-speaking users have significantly benefited from these advancements, the Arabic world faces distinct challenges in developing Arabic-specific LLMs. Arabic, one of the languages spoken most widely around the world, serves more than 422 million native speakers in 27 countries and is deeply rooted in a rich linguistic and cultural heritage. Developing Arabic LLMs (ALLMs) presents an unparalleled opportunity to bridge technological gaps and empower communities. The journey of ALLMs has been both fascinating and complex, evolving from rudimentary text-processing systems to sophisticated AI-driven models. This article explores the trajectory of ALLMs, from their inception to the present day, highlighting the efforts to evaluate these models through benchmarks and public leaderboards.

The 2023 Türkiye-Syria earthquakes, also known as the 2023 Kahramanmaraş earthquakes, were two catastrophic events that struck nine hours apart on February 6, 2023, with epicenters in the Pazarcık and Elbistan districts of Kahramanmaraş, and magnitudes of 7.8 Mw and 7.5 Mw, respectively (see Figure 1).

This week has been an eventful one for America's public health agency. Two former leaders of the US Centers for Disease Control and Prevention explained why they suddenly departed in a Senate hearing. They also described how CDC employees are being instructed to turn their backs on scientific evidence. They painted a picture of a health agency in turmoil--and at risk of harming the people it is meant to serve. And, just hours afterwards, a panel of CDC advisers voted to stop recommending the MMRV vaccine for children under four. This article first appeared in The Checkup, MIT Technology Review's weekly biotech newsletter.

The skeletal remains of a man who went missing six years ago were found by a walker in a secluded area in south Wales, an inquest has heard. Jordan Moray, from Cwmbach, near Aberdare in Rhondda Cynon Taf, was reported missing from his flat with his games console still running and mobile phone on charge in July 2019. Despite extensive police searches, his remains were not found until 29 August 2025 . On Friday, an inquest at Pontypridd Coroner's Court heard the discovery was made in a remote area near Merthyr Tydfil. South Wales Police previously said it had received a report of human remains near the Llwyn-on Reservoir in Bannau Brycheiniog National Park, also known as the Brecon Beacons .

A drone crashed into a hotel in the southern Israeli city of Eilat on Thursday, causing a fire but no casualties, authorities said. Yemen's Houthi group, who have been firing drones and missiles in solidarity with Palestinians in Gaza, claimed responsibility for the attack. Palestinians turn to the sea to flee Israel's bombardment Trump says US wants Afghanistan's Bagram Air Base back from Taliban What did Jimmy Kimmel say about Charlie Kirk's killing?

Kim Jong Un declares AI military drone development a'top priority' North Korea's Supreme Leader Kim Jong Un has said the use of artificial intelligence is a "top priority" in modernising his country's increasingly sophisticated weapons technology and building up drone capabilities, state media reports. During a visit to the Unmanned Aeronautical Technology Complex in the capital Pyongyang on Thursday, Kim presided over performance tests of multipurpose drones and unmanned surveillance vehicles, North Korea's Korean Central News Agency (KCNA) said on Friday. Kim also called for "expanding and strengthening the serial production capacity of drones". The visit to the aeronautical complex comes just a week after Kim oversaw another test of a new solid-fuel rocket engine designed for intercontinental ballistic missiles, which he hailed as a "significant" expansion of Pyongyang's nuclear capabilities. North Korea's military power includes nuclear-armed ballistic and cruise missiles, an increasing stockpile of nuclear weapons and a nascent spy satellite programme, according to the United States Defense Intelligence Agency (DIA).

ChatGPT was used'to help scammers do their thing' at Asia fraud compound ChatGPT owner OpenAI says it actively works to identify and disrupt scam-related misuse of ChatGPT." Duncan Okindo says he was lured to Southeast Asia last year by the promise of a customer service job in Thailand. Instead, he ended up spending four months in a scam compound on the lawless Myanmar-Thai border, where he saw first-hand how criminal groups are at scale. Okindo, 26, says he was struggling to find a job as the breadwinner for his family in his native Kenya when a local recruitment agency promised him work in Bangkok. The flight was his first trip overseas. On landing, he says, he was abducted at the airport and spirited across the border, into the notorious KK Park complex, guarded by heavily armed men and fortified like it was meant for war."

The proliferation of artificial intelligence (AI) systems, from virtual assistants [Kepuska and Bohouta, 2018] to recommendation engines [Gomez-Uribe and Hunt, 2015] and autonomous vehicles [Shladover, 2018], has reshaped human-machine interaction. Y et, African languages, with over 2,000 spoken across the continent [Eberhard et al., 2023], remain severely underrepresented in NLP due to their low-resource status [Ahia and Boakye, 2023, Nekoto et al., 2020]. This exclusion risks exacerbating the digital divide, limiting access to AI-driven services in critical domains like education, healthcare, and governance [Ndichu et al., 2024, Joshi et al., 2020]. Shona, a Bantu language spoken by millions in Zimbabwe and southern Zambia, exemplifies this challenge. Existing Shona corpora primarily consist of formal texts, such as news articles or religious documents [Eberhard et al., 2023], while everyday communication, particularly among younger speakers, is dominated by slang, code-mixing with English, and informal expressions [Eisenstein, 2013]. Standard NLP models, trained on formal data, struggle to process these dynamic linguistic patterns, hindering the development of culturally relevant conversational AI.

Malnutrition among pregnant women is a major public health challenge in Ethiopia, increasing the risk of adverse maternal and neonatal outcomes. Traditional statistical approaches often fail to capture the complex and multidimensional determinants of nutritional status. This study develops a predictive model using ensemble machine learning techniques, leveraging data from the Ethiopian Demographic and Health Survey (2005-2020), comprising 18,108 records with 30 socio-demographic and health attributes. Data preprocessing included handling missing values, normalization, and balancing with SMOTE, followed by feature selection to identify key predictors. Several supervised ensemble algorithms including XGBoost, Random Forest, CatBoost, and AdaBoost were applied to classify nutritional status. Among them, the Random Forest model achieved the best performance, classifying women into four categories (normal, moderate malnutrition, severe malnutrition, and overnutrition) with 97.87% accuracy, 97.88% precision, 97.87% recall, 97.87% F1-score, and 99.86% ROC AUC. These findings demonstrate the effectiveness of ensemble learning in capturing hidden patterns from complex datasets and provide timely insights for early detection of nutritional risks. The results offer practical implications for healthcare providers, policymakers, and researchers, supporting data-driven strategies to improve maternal nutrition and health outcomes in Ethiopia.

This paper discusses our exploration of different data-efficient and parameter-efficient approaches to Arabic Dialect Identification (ADI). In particular, we investigate various soft-prompting strategies, including prefix-tuning, prompt-tuning, P-tuning, and P-tuning V2, as well as LoRA reparameterizations. For the data-efficient strategy, we analyze hard prompting with zero-shot and few-shot inferences to analyze the dialect identification capabilities of Large Language Models (LLMs). For the parameter-efficient PEFT approaches, we conducted our experiments using Arabic-specific encoder models on several major datasets. We also analyzed the n-shot inferences on open-source decoder-only models, a general multilingual model (Phi-3.5), and an Arabic-specific one(SILMA). We observed that the LLMs generally struggle to differentiate the dialectal nuances in the few-shot or zero-shot setups. The soft-prompted encoder variants perform better, while the LoRA-based fine-tuned models perform best, even surpassing full fine-tuning.