Latest AI News

Get the fastest, most accurate AI news every morning

Total articles: 13,568|Added todayToday: 69

Creative Ideaz Launches Birmingham AI Bootcamp

Creative Ideaz launched a half-day Birmingham AI Bootcamp for business owners and senior leaders.
UK business AI adoption reached 32% in 2026, up from 16% in 2025, the government survey found.
Survey respondents cited skills barriers, while 81% of AI users or planners reported human oversight of outputs.

lfpress.comRead more · 2 min →

Yesterday'sInfra

South Korea and AMD Discuss AI Chip Cooperation

South Korea’s science ministry and AMD held a committee meeting on their AI chip cooperation agreement.
They discussed linking domestic NPUs with AMD CPUs and GPUs, and plans for computing infrastructure and an AI research center.
AMD held its Advancing AI event in Seoul for the first time in South Korea, with more than 1,300 preregistrations.

Seoul Economic DailyRead more · 1 min →

Yesterday'sInfra

AMD Plans Seoul AI Test Cluster for Next Spring

AMD, FuriosaAI and MangoBoost plan to build an AI infrastructure test cluster in Seoul next spring.
The cluster will combine CPUs, GPUs, an NPU and a DPU to assess infrastructure efficiency and potential uses.
AMD also plans to discuss establishing multiple Centers of Excellence in South Korea next year.

Money TodayRead more · 1 min →

Yesterday'sCoding

Ricursive Releases Ripple-65B Preview

Ricursive released Ripple-65B Preview, reporting 80.52 on TerminalBench 2.1 for U.S. open-weight agentic coding.
The company says training and full R&D cost $32 million; the model used 12 trillion tokens and supports a 1 million-token context.
Ricursive reports 1.4 million reinforcement-learning rollouts and up to 16 times better code compute efficiency at 4 billion parameters.

ricursive.comRead more · 2 min →

Yesterday'sHealthcare

Review Assesses AI Evidence in Nephrology

A 2026 Frontiers in Nephrology review assesses AI evidence across kidney care and clinical readiness
The review searched PubMed for studies from 2020 to 2025, retaining two foundational pre-2020 studies
Prospective evaluation and clinical utility evidence remain limited across many nephrology AI applications

Semantic ScholarRead more · 1 min →

Yesterday'sInfra

IBM Japan Expands AI-Led Business Transformation Service

IBM Japan is fully launching IBM Enterprise Advantage, a business transformation service staffed by expert teams including Forward Deployed Engineers.
Its Forward Deployed Unit has six roles and uses six AI agents for business analysis and design.
At Heineken, four local subsidiaries are operating the service in production; IBM Japan aims to begin 20 client engagements by the end of 2026.

Impress DIGITAL XRead more · 2 min →

Yesterday'sPolicy

Agencies Urged to Audit AI Data Readiness

Government Technology contributor calls for data audits before agencies put AI systems into production
Michigan review found a 93 percent error rate across more than 20,000 automated unemployment fraud determinations
Three audits examine data sources, prediction targets and safeguards, with named owners and a launch readiness gate

GovTech AIRead more · 2 min →

Trending Keywords

Yesterday'sInfra

Monsterlab Launches TakumiRebuild AI

Monsterlab launched TakumiRebuild AI to turn skilled workers’ judgments into rules.
A logistics placement proof of concept was built in 5 weeks, raising judgment agreement from about 50% to about 85%.
The proof of concept cut skilled workers’ labor by about one-third and supports reviewing and updating rules.

Impress DIGITAL XRead more · 1 min →

Yesterday'sInfra

HPE and AMD Outline Sovereign AI Stakes

HPE and AMD’s sponsored post describes sovereign AI investment as a geopolitical and economic priority for governments and businesses.
Bain cites billions in government spending; McKinsey projects a $600 billion sovereign AI market by 2030.
The post cites compliance, infrastructure security and operating costs as trade-offs for companies navigating national AI systems.

CIO.comRead more · 1 min →

Yesterday'sPolicy

Midterm Results Raise AI Policy Questions

Democrats swept the House and Senate as voters rejected data centers in 40 of 41 midterm ballot measures
Analysts say election results could increase AI oversight, state enforcement efforts and congressional investigations
Nearly one-third of expected 2028 data center additions are at risk, UBS said

CIO DiveRead more · 2 min →

Yesterday's

ToolSpace Builds Harder Agent Benchmarks

Researchers introduced ToolSpace, a tool-centric method for synthesizing environments for long-horizon agents.
Seven models dropped up to 38.2 points on harder AutomationBench and τ³-Bench Banking Pro versions.
Across seven benchmarks, Qwen3.6-35B-A3B rose from 23.0 to 41.0, or 43.2 with reinforcement learning.

HuggingFaceRead more · 1 min →

Yesterday's

Researchers Introduce VidUEU-Agent Video Data Pipeline

Researchers introduce VidUEU-Agent to synthesize multimodal instruction data from raw videos
Four-stage pipeline iterates on data distribution using downstream training and evaluation feedback
VidUEU-100K contains 100,000 samples from 8,268 videos across nine task paradigms

HuggingFaceRead more · 1 min →

Yesterday's

U Mobile Deploys Agentic Telecom Helpdesk

U Mobile built AiDaH with AWS and Axrail on Amazon Bedrock AgentCore for telecom operations.
AiDaH covers over 150 scenarios across six divisions; ticket volumes fell by up to 84.5 percent in key categories.
Eight specialist agents use three-tier request routing, with roughly 70 percent resolved by keyword matching without a model.

AWS ML BlogRead more · 2 min →

Yesterday's

YANAMI Tests Multimodal Models Across Dialogue Turns

Researchers publish YANAMI, a dataset and benchmark for multimodal conversations that change across turns
YANAMI-400K contains 400K dialogues, more than 2M images and 2.6M dialogue turns
Fine-tuned BAGEL improved image editing 10.3%, generation 2.0% and a third reported task 13.23%

HuggingFaceRead more · 1 min →

Yesterday'sInfra

Texas Election Pressures AI-Linked Stocks

Democrat Gina Hinojosa won Texas’s governor race after campaigning for an immediate data center moratorium
Wells Fargo estimates Texas holds 27% of planned U.S. data center gigawatts; Georgia has 7%, Ohio and Pennsylvania 5% each
SpaceX fell 4% and Tesla 5% as several AI-linked stocks slid after Democrats won both houses of Congress

CNBCRead more · 1 min →

Yesterday'sHealthcare

Pinal Designs Protein Candidates From Text

Westlake University researchers developed Pinal to design candidate proteins from natural-language descriptions of their functions.
Pinal training used 135 million protein sequences to build 1.7 billion protein-text pairs.
One of 29 tested fluorescent candidates worked; five of 20 H-protein candidates worked, with one reaching 1.7 times output.

phys.orgRead more · 1 min →

Yesterday'sInfra

OpenAI Restrains Public Comments on China Chip Controls

The Information reports OpenAI has limited public comments on China chip controls amid concern about worsening ties with Nvidia.
After an internal question from Vice President Noam Brown, policy staff directed China export-control discussions to global policy officials.
OpenAI supports strong controls but is cautious on China-specific rules; its Jalapeño chip is planned for use in the United States and allied countries.

AI Times KoreaRead more · 2 min →

Yesterday'sSafety & Ethics

EU Adviser Flags Copyright Risks in Gemini Article Reuse

An EU court adviser said Gemini’s reuse of news articles may infringe copyright under certain conditions.
Like Company alleged Google used its articles to train Gemini and reproduce them in answers from June 2023 to May 2024.
The adviser’s opinion is nonbinding; a Hungarian court will make the final ruling.

AI Times KoreaRead more · 1 min →

Yesterday'sHealthcare

FDA Receives 500 Comments on Medtech AI Rules

FDA received more than 500 comments by mid-October on regulating generative AI in medical devices.
AdvaMed favored existing FDA rules, while physicians and patients sought stronger safeguards and postmarket monitoring.
Oura, Generation Patient and Stanford’s Brainstorm lab raised questions about wearables, companion bots and safety testing.

Healthcare DiveRead more · 2 min →

Yesterday's

AI Firms Compete for Forward-Deployed Engineers

WSJ reports top forward-deployed engineers can earn up to $1.45 million in annual total compensation.
The United States has about 17,000 forward-deployed engineers, including just 2,000 at the top level; about 80% have worked at Palantir.
Target and Barclays are hiring their own teams, while Google and Accenture plan to train up to 1,000 engineers together.

AI Times KoreaRead more · 1 min →

Trending Keywords

Last 7 Days