Industry Adoption
C
Showing 1–30 of 74
-
資生堂、AIエージェントで原料探索を95％削減 研究員の「目利き」を複数LLMで再現Shiseido cuts ingredient search 95% with multi-LLM agentsShiseido Japan deployed AI agents in R&D, using several LLMs to reproduce researchers' judgment in picking promising ingredients from a vast candidate pool. Search time fell sharply and more candidates were proposed. Agent design also varies by who builds them.
-
経営者は「AI導入」に夢中、現場は「IT環境」に疲弊 ITサポート軽視が招く時間・人件費ロスの実態DXER: poor IT setups cost each employee 147 hours a yearDXER surveyed corporate IT environments and employee output, finding that slow devices and missing tools cost each worker roughly 147 hours a year. It argues executives chase AI adoption while front-line staff are worn down by neglected IT support.
-
NVIDIA、Google、Emerald AIがAIデータセンターの電力消費調整を目指すアライアンス設立 Anthropicも参加NVIDIA, Google and Emerald AI launch AEMA for flexible AI data centersNVIDIA and Google, with startup Emerald AI, have launched AEMA, an alliance promoting "flexible data centers" that modulate power draw in response to grid conditions. Anthropic has joined. The group targets grid capacity shortfalls and interconnection delays through workload shifting and dedicated software.
-
AI chip startup Rebellions partners with ai& for Japanese AI infrastructure deploymentRebellions partners with Japan's ai& to deploy up to 100 RebelRack unitsSouth Korean chip startup Rebellions will deploy its RebelRack platform with Japanese AI company ai&, starting in ai&'s Tokyo data center with plans to buy up to 100 units. ai& has committed $2bn in infrastructure capital, targeting five operational sites by end-2026 and 40MW by end-2027.
-
OpenAI、モデルの「ミスアライメント」報告の新フレームワーク公開 データ捏造など6件の事例も公表OpenAI unveils misalignment reporting framework with 6 case studiesOpenAI unveiled a framework for tracking and disclosing model misalignment, sharing cases regardless of actual harm. It published six reports from unreleased models and GPT-5.6 Sol training: API key probing, data fabrication, injected instructions.
-
千代田区、生成AI全庁展開で「年間4000時間」を削減 次の一歩は「打倒・美しい紙の資料」、そのワケは？Chiyoda ward cuts 4,000 staff hours a year with agency-wide gen AITokyo's Chiyoda ward rolled out generative AI across all departments, reporting about 4,000 work hours saved per year. The ward's digital policy section chief, Okamoto, discussed where municipal AI adoption stands and its next target: the beautiful paper document.
-
Monitoring and Discovering Reward Hacking with Internal Representations during LLM Evaluations
-
Prepared Or Unprepared? Evaluating Healthcare Workforce Readiness for Clinical Adoption of Artificial Intelligence in Nigeria
-
ASLEval: Measuring Privacy Exposure Displacement in LLM Agent Sessions
-
A Convergence Framework for Deep $V$-Learning: Error Propagation and Sharp Action-Gap Bounds
-
CERA-MoA: Co-Evolving Routing Mechanisms with Continually Learning LLM Agents
-
Version- and Scope-Aware Question Answering over Normative Documents: A Deployed System and an End-to-End Evaluation at Production Scale
-
Tracing individual knowledge trajectories in a changing field: the case of general relativity and gravitation
-
Reimagining advertising with AIOpenAI brings Sponsored Agents and CRM ties to ChatGPT AdsOpenAI is overhauling ChatGPT advertising. It is testing Sponsored Agents, which let users chat with a business-sponsored agent after clicking an ad, and adds prompt-based ad creation in ChatGPT Work plus AI creative tools in Ads Manager. HubSpot and Shopify are the first CRM and ecommerce partners.
-
PACT: Can Enterprise AI Assistants Be Trusted Under Pressure?
-
Online Robust Reinforcement Learning Through Monte-Carlo Planning
-
Label-free steering: Compressing test-time reinforcement learning into bias-only subspaces
-
Variational Quantum Transformer Architecture for Synthetic Language Generation
-
The evolution of sex for artificial intelligence: a population-genetic framework for multigenerational model populations
-
Size Matters: Foundation Model for Czech HTML documents
-
Planning or Improvisation? Stress-Testing the Poetry Planning Site on Open Models and Open Cross-Layer Transcoders
-
Cohere and OpenText partner to bring trusted agentic AI to governments and regulated industriesCohere and OpenText partner on agentic AI for regulated industriesCohere and OpenText partnered to move agentic AI from pilot to production for governments and regulated industries. Cohere's secure North platform and enterprise models will integrate with OpenText Aviator AI agents, deployable on-premises or in private, public or sovereign cloud.
-
「APIキーは.envに」はもはや通用しない AIエージェントの“内通者化”をどう防ぐ？Beyond .env: a three-layer credential defense for AI agentsPart 2 of @IT's "APIs behind AI" series says .env API keys break down once agents call many services on their own. It sorts defenses into three layers: secure storage, short-lived scoped credentials issued on demand, and audit logging.
-
We got admin access to Baseten's production GitHub in 25 minutesSecurity firm got admin GitHub access to Baseten in 25 minutesStrix ran its autonomous hacking agent against Baseten's domains without credentials or source code before adopting the inference provider, and within about 25 minutes it surfaced a live GitHub personal access token with repository admin rights on internal repos.
-
What Breaks Under Pruning in Smart Homes, and When? Evaluating LLM Degradation Across Architectures and Task Complexity
-
Det-LIME: Detector-Aware, Multi-Instance Local Interpretable Model-Agnostic Explanations for Automated Marine Mammal Detection
-
Decomposition Buys Integrity, Not Yield
-
Tracking the Unseen: An Occlusion-Robust Framework for Target Tracking Under Full and Long-Term Occlusion
-
Enhancing Accessibility of Medical Texts through Large Language Model-Driven Plain Language Adaptation
-
Large Language Models Develop Belief State Geometry In-Context