No news signals yet.
Researchers analyzed 14,419 self-published books on Amazon from 2023 to 2026 and found that books with substantial AI-produced content make up a large share of the catalog. These books win a growing s…
A study comparing Canadian and South Korean university students found significant cross-cultural differences in perceptions of generative AI use, with Canadian students more likely to view it as uneth…
Researchers used large language models from OpenAI, Anthropic, and Google to simulate freight matching markets with 50 shipper agents. The simulations revealed concentration risks, with a single carri…
Researchers present a regression-based approach to Arabic dialect geolocation, predicting speaker origin as continuous latitude-longitude coordinates with a median localization error of 481.2 km. The…
A qualitative study of 25 construction artisans in North-East Italy examined their perspectives on technical innovation. The study identified 15 premises underlying their views, grouped around four is…
Researchers develop Eutopia, a lending-process simulator, to study long-term fairness in AI-driven decision makers, finding that learning with performative dynamics leads to better efficiency and equi…
Researchers propose using Clinical Pathways as safety specifications for Physical AI in hospital wards, integrating wearable sensors and robotic components. A Runtime Safety Monitor evaluates signals…
AI integration in talent management helps organizations design strategies for talent retention and engagement. Predictive models and personalized career planning can address retention issues and impro…
Researchers analyze attributions of consciousness to AI chatbots, finding little scientific evidence to support these claims. A multidimensional taxonomy is developed to categorize attitudes towards A…
IBM reported a disappointing quarter with revenue of $17.2 billion, missing Wall Street's expectations. CEO Arvind Krishna attributed the 42% drop in mainframe sales to customers prioritizing other ha…
Synthesia has launched Roleplay Sessions, an AI-powered training platform where employees practice high-stakes conversations with an avatar. The British startup's new product aims to help enterprises…
Researchers studied Large Language Models' persuasive potential using Jürgen Habermas' Theory of Communicative Action. They found that LLMs effectively convey illocutionary intent, often more so than…
Researchers identify Safety Drift and Operational Hallucination in large language models, where initial alignment degrades over time, leading to constraint-violating actions. These phenomena are obser…
Researchers introduce Interpretive Knowledge Nodes and Citation Compression Layers to analyze citation pathways in the AI era. AI transforms the cost structure of producing citable knowledge intermedi…
Researchers introduce a mathematical framework to study AI value alignment with evolving social norms, highlighting risks like value lock-in and normative mode collapse. The framework uses social phys…
Researchers analyzed 2,611 Spanish mobile apps using large language models, finding discrepancies between declared and observed privacy practices, especially in public-sector apps. The study reveals l…
Researchers address gaps in the concept of loss of control in AI, establishing a definition of control based on setting and getting goals, and discussing how control can be lost, including through AI…
Researchers propose MR-ConceptGCN, a multi-relational graph convolutional network approach for sequential learner modeling, combining personal knowledge graphs and pre-trained language models. The met…
Researchers propose a Two TimeScale Reinforcement Learning framework for delivery drone-based participatory sensing in dynamic environments. The framework, called TSRL, separates decision-making into…
Researchers analyzed 18,797 work tasks in 124 economies, finding that feasible automation depends on task content and country-level conditions. The exposed share of tasks ranges from 3.3% to 61.6%, va…
Researchers compared TF-IDF + Logistic Regression and fine-tuned DistilBERT for phishing email detection, finding both exceeded 98% accuracy on clean data but dropped to 64.00% and 63.64% under advers…
Researchers argue that modern AI systems face hidden safety-critical challenges, including quieter failures that are plausible and distributed across components. A proposed five-layer framework identi…
Researchers evaluated 41,331 summaries from 13,777 articles, finding AI-powered browsers like Google Chrome and Microsoft Edge to be broadly accurate and effective at reducing political bias and negat…
Researchers compared GPT-generated surveys to human-designed ones across three social domains. The GPT-generated surveys captured dominant attitudinal divisions, but showed differences in belief struc…
African governments are prioritizing AI opportunities for economic transformation, but devote little attention to AI safety, according to an analysis of national AI strategies and public statements. T…
Researchers introduce digital statecraft as a concept to govern the digital substrate of states, comprising data, algorithms, and infrastructure. The concept operates on two dimensions, requiring tech…
Researchers analyzed NFL tweets from 2011-2014, finding sentiment for teams is positive before games, drops at kickoff, and diverges after halftime, with winning teams' sentiment increasing and losing…
Anthropic and Physical Intelligence held acquisition talks this spring, according to The Information, despite Physical Intelligence's CEO denying the rumor. Physical Intelligence, co-founded by Lachy…
Dimension Capital, a New York firm, has announced an $800 million fund, 60% larger than its previous $500 million fund. Founded by Zavian Dar, Adam Goulburn, and Nan Li, the firm invests in companies…
Meta is piloting StoryKit, an AI storytelling app, in select countries to create personalized children's stories with custom characters and lessons. The app uses AI safety filters and is available to…
