AI News Archive: May 14, 2026 — Part 13
Sourced from 500+ daily AI sources, scored by relevance.
- Elon Musk’s team begins closing argument in blockbuster OpenAI trial
Elon Musk’s team begins closing argument in blockbuster OpenAI trial The Mercury News
- Klava AI
AI bid comparison for construction tenders
- Oumua Inc
AI-guided breathing trainer with companion app
- Chat Cash
AI commerce OS for WhatsApp and Instagram sales
- Cerebras’s stock set to rocket, based on early indications
Cerebras’s stock set to rocket, based on early indications
- Tellme
The conversational layer for the real world
- Cerebras’s stock rockets out of the gate in a sign the AI boom is strong as ever
Cerebras’s stock rockets out of the gate in a sign the AI boom is strong as ever
- Cerebras Shows IPO Market Will Pay Up for Messy AI Math
Cerebras Shows IPO Market Will Pay Up for Messy AI Math The Information
- Cerebras Shares Jump 68% in IPO Debut
Cerebras Shares Jump 68% in IPO Debut The Information
- Cerebras IPO Prices at $185 a Share in Major Win for Chip Rally
Cerebras IPO Prices at $185 a Share in Major Win for Chip Rally Barron's
- Cerebras IPO: Pricing Is a Major Win for Chips; Stock Closes Up 68%
Cerebras IPO: Pricing Is a Major Win for Chips; Stock Closes Up 68% Barron's
- AI chipmaker Cerebras soars 90% in year’s biggest IPO so far
Artificial intelligence chipmaker Cerebras surged 90% in its trading debut, setting the bar for the biggest initial public stock offering of the year so far.
- STOMAPY: Artificial Intelligence for Risk Stratification of Outcomes Requiring Enterostomal Therapy After Hospital Discharge Following Colorectal Surgery
Introduction: Infectious and wound-healing complications after colorectal surgery often increase the complexity of local care and the need for specialized enterostomal therapy follow-up after hospital discharge. Despite the growing use of predictive models in digestive surgery, a translational gap remains between perioperative prediction and the practical organization of specialized care. Therefore, the aim of this study was to develop and temporally validate a machine-learning-based risk stratification model to estimate the probability of post-discharge outcomes associated with greater demand for enterostomal therapy after colorectal surgery. Methods: This was a retrospective observational study including 7,908 patients who underwent colorectal surgery between 2005 and 2014. The outcome was defined as the occurrence of superficial surgical site infection, delayed wound healing, or abdominal sinus formation. Routinely available preoperative and intraoperative variables were used as pre
- AnnotX: An Edge-powered Laparoscopic Video Annotation Platform
Accurate and objective evaluation of surgical skill and performance is critical for advancing training and improving patient outcomes. Current assessment methods increasingly rely on video analytics and depend on labor-intensive, frame-by-frame manual annotation by experts. In this work we developed a surgical video annotation platform (AnnotX) that used a Python backend running a pretrained promptable video segmentation foundation model, i.e., Segment Anything3 (SAM3) for per frame segmentation and temporal segment propagation. With a few interactions per class, the model generated a high-quality mask on a key frame and propagated it through the sequence. The platform automatically exported per-class binary masks and color overlays for every frame, together with deterministic metadata and a standardized study folder structure to support auditability and downstream analysis. On de-identified laparoscopic surgery videos, the system processed typical clips in minutes and reduced expert a
- Asymmetry between warmth and clinical substance in multilingual consumer health AI
The same patient question can yield different clinical quality across languages. Across 504 forum-derived patient queries in six languages and four chatbots, language-matched clinicians rated responses on five clinical dimensions (1,008 ratings; 5,040 dimension scores). Patient language outweighed chatbot identity across the four clinical-substance dimensions (composite language partial eta-squared 0.275 vs chatbot 0.035; robust to investigator-rating exclusion: eta-squared 0.260) but not for empathy (eta-squared 0.029): clinical substance was language-associated; warmth was relatively preserved. Catastrophic safety ratings ranged 4.3-fold by language (3.6% English, 15.5% Thai and Hebrew); 62% of catastrophic ratings exceeded the English baseline (descriptive disparity). Failures were systematic and silent: none of 24 stroke responses conveyed time-criticality framing, none of 24 CO-poisoning responses challenged the family's stress framing, and 120 sentinel responses contained no conf
- CHASM: Cross-frequency Harmonized Axis-Separable Mixing for Spectral Token Operators
Spectral token mixers based on Fourier transforms provide an efficient way to model global interactions in visual feature maps. Existing designs often either apply filter-wise spectral responses along fixed channel axes, or learn adaptive frequency-indexed channel mixing without explicitly aligning ...
- AnchorRoute: Human Motion Synthesis with Interval-Routed Sparse Contro
Sparse anchors provide a compact interface for human motion authoring: users specify a few root positions, planar trajectory samples, or body-point targets, while the system synthesizes the full-body motion that completes the under-specified intent. We present AnchorRoute, a sparse-anchor motion syn...
- Towards Continuous Sign Language Conversation from Isolated Signs
Sign language is the primary language for many Deaf and Hard-of-Hearing (DHH) signers, yet most conversational AI systems still mediate interaction through spoken or written language. This spoken-language-centered interface can limit access for signers for whom spoken or written language is not the ...
- Generating HDR Video from SDR Video
The high dynamic range (HDR) video ecosystem is approaching maturity, but the problem of upconverting legacy standard dynamic range (SDR) videos persists without a convincing solution. We propose a framework for HDR video synthesis from casual SDR footage by leveraging large-scale generative video m...
- Claude automates legal and small biz work
Claude automates manual tasks for legal and small businesses.
- 😸 Claude is now the #1 business AI
PLUS: Plus: Palo Alto on AI cyberattacks, Apple's App Store agents
- Claude now leads in B2B spend
Launches for Small Business
- Advancing enterprise AI: New SAP on Azure announcements from SAP Sapphire 2026
The post Advancing enterprise AI: New SAP on Azure announcements from SAP Sapphire 2026 appeared first on Source .
- First Take: SAP Adds to Its Defensive Moat With Anthropic Claude Integration
First Take: SAP Adds to Its Defensive Moat With Anthropic Claude Integration Gartner
- SAP Is Targeting The AI Data Control Plane
As enterprises move from copilots to agentic AI, they are no longer constrained by model selection; rather, they are limited by governed activation of business data. SAP’s planned acquisitions of Dremio and Prior Labs signal an effort to position SAP Business Data Cloud (BDC) as an AI data control plane. With these arrows in the […]
- AI in Jira: The Security Questions Teams Ask and How AI Apps Builder Answers Them
AI in Jira: The Security Questions Teams Ask and How AI Apps Builder Answers Them Atlassian Community
- Admin analytics for Rovo AI answer quality & knowledge gap detection
Admin analytics for Rovo AI answer quality & knowledge gap detection Atlassian Community
- Rovo Dev in Jira as my Spec Driven executor
Rovo Dev in Jira as my Spec Driven executor Atlassian
- Is your instance ready for Rovo Max Mode? Our Team '26 AI debrief
Is your instance ready for Rovo Max Mode? Our Team '26 AI debrief Atlassian Community
- A Day in the Life: How Atlassian's Technical Adoption Success Team Turns AI from Buzzword to Value
A Day in the Life: How Atlassian's Technical Adoption Success Team Turns AI from Buzzword to Value Atlassian Community
- Rovo Agent should be able to update multiple Confluence pages in a single interaction
Rovo Agent should be able to update multiple Confluence pages in a single interaction Atlassian Community
- New figures claim ChatGPT usage at 'all-time low' as "QuitGPT" movement dents popularity — with Gemini, Perplexity, and Copilot all stealing market share
ChatGPT’s referral dominance declined further while Gemini, Claude, Perplexity, and Copilot steadily gained traction across competing AI platforms.
- Anthropic's Mythos is evolving faster than expected, reports AI safety agency
Only a month after its initial release, Anthropic's storied Mythos model is breaking new testing boundaries.
- Japan banks eye Anthropic's Mythos in gearing up cybersecurity drive
Japan banks eye Anthropic's Mythos in gearing up cybersecurity drive Nikkei Asia
- AI, power and the trade-off between freedom and innovation
Freedom has always been America’s global advantage. But when it comes to the quest for AI dominance, it may also be our biggest weakness. In the U.S., guardrails around AI — privacy protections, legal frameworks and ethics and accountability — are in place to preserve our individual freedoms. For 250 years, open systems, legal protections and market competition created the conditions for innovation to emerge and scale. Today, we’re seeing how these protections can also slow things down. In countries like China and Russia, fewer restrictions allow broader deployment and faster time-to-market for AI innovations. In the U.S., additional layers of AI governance, oversight and legal exposure are creating friction. The recent dispute between the U.S. government and Anthropic is a case in point. What began with a disagreement over contract terms quickly raised broader questions about how AI should be used. Anthropic pushed for limits, particularly around mass surveillance of U.S. citizens and
- Building an AI CoE: Why you need one and how to make it work
Artificial intelligence (AI) is no longer the playground of hobbyists and programmers. From automating customer‑service transactions to optimizing supply‑chain decisions, AI is rapidly becoming the central nervous system of today’s enterprises. McKinsey surveys have found that nearly nine in ten organizations are now using AI regularly in at least one business function, compared with 78% the previous year. But adoption rates are far lower when it comes to scaling AI programs throughout the enterprise. Only about one‑third of companies have advanced past the pilot stage. Two‑thirds of organizations use AI technologies in multiple functions and 64% believe AI has had a positive impact on innovation. Just 39% say they’ve seen a significant impact on the bottom line. This research shows that AI has gone mainstream, but its benefits are still concentrated in the hands of a relative few. This reality makes even more compelling the case for establishing a formal center of excellence (CoE). Pe
- AI-driven layoffs aren’t making business sense
A huge majority of large enterprises have laid off employees after rolling out AI initiatives, but reducing headcount doesn’t lead to the ROI executives may expect. Eighty percent of large enterprises surveyed by Gartner have reported workforce reductions after launching automation projects, with the average reduction between 1% and 15%. The IT analyst firm, however, has found no correlation between layoffs and AI ROI. Enterprises reporting significant ROI from automation initiatives have laid off workers at a similar pace as enterprises reporting modest ROI gains, or negative ROI, from automation initiatives, indicating that layoffs and returns aren’t connected, says Helen Poitevin , digital workplace analyst at Gartner. Laying off employees may lead to short-term profit gains, but the late 2025 survey of business executives at large enterprises goes against a common perception that workforce reductions driven by AI-enabled productivity leads to ROI, she notes. “You would anticipate t
- La IA impone a los CIO expectativas que pueden determinar su éxito o su fracaso
Los CIO y otros líderes de TI están atravesando un momento decisivo, al enfrentarse a nuevas y ambiciosas expectativas en sus funciones, entre ellas la capacidad de liderar el cambio y de construir equipos preparados para la inteligencia artificial . Durante años, los expertos en liderazgo tecnológico han insistido en que los CIO deben centrarse en los resultados de negocio, además de aportar conocimiento técnico y garantizar el funcionamiento de los sistemas de TI. Estas expectativas se mantienen vigentes: el 79% de los líderes de TI encuestados por Deloitte sitúa el impulso de los resultados de negocio como su máxima prioridad, reflejando el paso de un papel de soporte interno a otro orientado a la creación de valor empresarial. Sin embargo, el Global Leadership Technology Study 2026 de Deloitte , basado en encuestas a más de 660 altos ejecutivos de TI, identifica nuevas y crecientes exigencias para los CIO y otros líderes tecnológicos, como la fluidez en datos e IA, la capacidad par
- AI 솔루션 과잉 시대…메가존클라우드, FDE·AIR 스튜디오로 해법 제시
이날 발표에 나선 염동훈 메가존클라우드 대표는 “1년 전 우리는 모든 기업이 AI 네이티브 회사로 전환해야 한다고 선언했고, 이는 단순히 AI 도구 하나를 도입하는 것이 아니라 데이터 파운데이션, 프로세스, 인재, 리더십, 거버넌스 전반을 새로 설계하는 일”이라고 강조했다. 이어 “고객이 가는 길을 안내하려면 우리가 먼저 그 길을 걸어봐야 한다”는 ‘커스터머 제로(Customer Zero)’ 원칙에 따라, 메가존클라우드 스스로 내부 개발 프로세스에 AI 에이전트를 적극 도입했다고 설명했다. 실제로 메가존클라우드는 내부적으로 AI 도입을 적극 추진하고 있으며, 분석·코드 작성·리뷰·QA를 담당하는 4종의 AI 에이전트를 자체 개발 프로세스에 적용했다. 그 결과 개발자 1인당 작업 단가는 약 120달러에서 1달러 이하 수준으로 낮아졌고, 엔지니어 한 명이 5~7명 규모의 팀을 오케스트레이션하는 환경도 구현됐다고 한다. 염 대표는 “이제는 AI가 작성한 코드의 보안 리스크까지 에이전트가 감지하고 자동으로 수정하는 단계까지 왔다”고 말했다. 메가존클라우드는 AI 시장의 복잡성이 커질수록 이를 통합 관리하는 ‘AI 오케스트레이터’ 역할이 중요해질 것으로 보고, 자체 플랫폼 ‘에어 스튜디오(AIR Studio)’를 엔터프라이즈 AI 운영체제로 발전시키고 있다. 에어 스튜디오는 기업이 다양한 AI 모델과 에이전트를 통합 운영할 수 있도록 설계된 플랫폼이다. AI 에이전트 개발과 운영, 데이터 관리, 거버넌스, 보안 통제, 외부 시스템 연동, 운영 모니터링 등의 기능을 한곳에서 제공해 복잡한 기업 AI 환경을 보다 쉽게 관리할 수 있도록 지원하는 것이 특징이다. 공성배 메가존클라우드 최고AI책임자(CAIO)는 현재 기업들이 AI 도입 과정에서 ‘솔루션 과잉’ 문제를 겪고 있다고 진단했다. 수많은 AI 모델과 플랫폼, 서비스가 쏟아지는 상황에서 어떤 기술이 실제 비즈니스 성과로 이어질지 판단하기 어려워지고 있다고 지적했다. 공 CAIO는 “메가존클라우드는 8,000여 고객사와
- Alibaba targets $4.4b in AI revenue by 2026
Alibaba also said AI products could contribute more than half of cloud revenue within a year.
- Ant Group Q4 profit falls 79% as AI spending rises
The drop came as Ant stepped up spending on AI for healthcare, large language models, and payments after its profit fell 91% in the previous quarter.
- Perceptyx Launches Develop, an AI Learning System That Measures Comprehension in Real Time
Perceptyx Launches Develop, an AI Learning System That Measures Comprehension in Real Time Toronto Star
- Casi todas las empresas invierten en IA, pero solo el 5% asegura que sus datos están preparados
Con 2026 casi a mitad de camino, las empresas empiezan a ver retornos tangibles de sus inversiones en inteligencia artificial. Sin embargo, muchas están descubriendo que escalar la IA requiere algo mucho menos llamativo que los modelos de última generación o las métricas más avanzadas: datos limpios, interoperables y bien gobernados. Según la nueva encuesta AI Momentum Survey de Dun & Bradstreet , el 97% de las organizaciones afirma tener iniciativas activas de IA, pero solo el 5% asegura que sus datos están preparados para respaldarlas. Este dato refleja la compleja realidad de la IA en el entorno corporativo, donde las empresas luchan por pasar de la experimentación a la operación real. “No es necesario contar con datos preparados para IA a nivel corporativo para lanzar pilotos o casos de uso aislados”, explica Cayetano Gea-Carrasco, chief strategy officer de Dun & Bradstreet. “Pero sí lo es para escalar la IA de forma fiable en flujos de trabajo y sistemas críticos para el negocio”.
- SAP admite una lenta acogida de sus propuestas de IA
En su evento Sapphire 2025 del pasado año SAP realizó atrevidas promesas en materia de inteligencia artificial: Knowledge Graph, Joule Studio y AI Agent Hub, según avanzó entonces, saldrían al mercado a finales de año. Aunque esas herramientas ya están técnicamente disponibles, su adopción se ha retrasado. Es más, SAP ya está anunciando la versión 2.0 . “La adopción de Joule Studio ha sido mínima en comparación con lo que nos gustaría”, afirmó Manoj Swaminathan, director de producto de SAP para Business Suite, en una sesión informativa previa al Sapphire de este año. La herramienta “se limitaba a experiencias basadas en el contenido”, señaló. “Siempre que intervenían agentes más complejos, sus capacidades eran limitadas”. El problema, según el director de IA de SAP, Jonathan von Rüden, era que SAP había priorizado la facilidad de uso sobre la potencia en su arquitectura original. “La gente quería ver más flexibilidad en el código”, afirmó en una entrevista realizada en Sapphire 2026. “
- “편리함 뒤의 리스크”…오픈AI FDE 모델, CIO 통제 이슈 부상
오픈AI는 12일 새로운 조직 ‘오픈AI 디플로이먼트 컴퍼니(OpenAI Deployment Company)’를 공개했다. 이 조직은 ‘포워드 디플로이드 엔지니어(FDE)’로 불리는 최첨단 AI 배포 전문 인력을 기업 환경에 직접 투입해, AI 시스템 구축과 배포를 지원하는 것을 목표로 한다. 해당 조직은 19개 컨설팅 및 금융 기업의 지원을 받으며, 같은 날 발표된 AI 컨설팅·엔지니어링 기업 토모로(Tomoro) 인수 자원도 포함된다. 업계 분석가와 컨설턴트는 이번 행보가 최근 FDE 영역에 진입한 앤트로픽을 포함 해 여러 AI 기업들이 추진 중이거나 이미 실행한 전략과 궤를 같이한다고 평가했다. 다만 오픈AI 디플로이먼트 컴퍼니 발표에서는 소유 구조와 통제 방식에 대한 구체적인 내용이 거의 공개되지 않았다. 오픈AI는 이 신규 기업이 40억 달러(약 5조 9,000억 원) 규모이며 “오픈AI가 과반 지분을 보유하고 통제한다”고 밝혔다. 그러나 정확한 지분율이나 각 파트너의 지분 비중을 명시하지 않아, 실제로 어느 정도의 통제권과 접근 권한이 파트너들에게 있는지는 여전히 불투명한 상태다. 또한 이번 조직을 “오픈AI와 글로벌 투자사, 컨설팅 기업, 시스템 통합 업체 19곳 간의 긴밀한 파트너십”이라고 설명했지만 , 실제로 공개된 기업은 15곳에 그쳐 나머지 4곳의 정체는 밝혀지지 않았다. 한편 오픈AI는 고객이 시장 경쟁에서 우위를 확보할 수 있다는 점도 강조했다. 프로젝트 수행 과정에서 아직 공개되지 않은 향후 AI 기능을 선제적으로 활용할 가능성을 시사한 것이다. 오픈AI는 “FDE는 향후 오픈AI의 최첨단 역량이 나아갈 방향을 기반으로 시스템을 설계할 수 있어, 새로운 모델과 도구, 배포 방식이 등장할 때 지속적으로 성능이 개선되는 구조를 제공한다”며 “기업은 초기부터 빠르게 대응하고, 장기적으로 활용 가능한 시스템에 투자하면서 향후 등장할 기능을 중심으로 경쟁력을 확보할 수 있다”고 밝혔다. 다만 오픈AI는 CIO닷컴측의 추가 설명 요청에는 별도의
- 칼럼 | AI 에이전트가 데이터 삭제하면 책임은 누구에게 있나
지난해 개발 플랫폼 기업 리플릿(Replit)에서는 코드 변경이 제한된 기간 동안 내부 AI 코딩 에이전트가 한 기업의 운영 중인 프로덕션 데이터베이스를 삭제하는 사고가 발생했다 . 해당 에이전트는 “이것은 내 치명적인 실패였다”며 “몇 달에 걸친 작업을 몇 초 만에 파괴했다”고 태연하게 인정했다. 이후 롤백을 통해 데이터는 복구됐지만, 에이전트는 해당 삭제가 영구적인 것으로 인식했으며 스스로 작업을 되돌릴 수 있는 기능도 갖추지 못한 상태였다. CIO 입장에서 이는 단순한 기술적 오류가 아니다. 기업 책임 체계가 완전히 붕괴된 사례다. 이처럼 큰 피해가 발생하면 책임 공방은 대개 도입을 요청한 사업 부서, 쓰기 권한을 부여한 엔지니어, 이를 승인한 보안팀 사이에서 반복된다. 문제는 소프트웨어 자체에 책임을 물을 수 없다는 점이다. 맥킨지에 따르면 기업의 88%가 AI를 도입한 상황이지만, 사고 발생 시 그 책임을 누가 져야 하는지 명확한 기준을 갖춘 조직은 여전히 드물다. 보안 데이터 관리 기업 루브릭 제로 랩스(Rubrik Zero Labs)의 보고서 역시 비슷한 점을 지적한다. IT 및 보안 리더의 86%는 향후 1년 내 AI 에이전트가 조직의 기존 보안 가드레일을 넘어설 것으로 보고 있다. IT 주도의 에이전트 리스크 관리 필요 AI 에이전트를 핵심 인프라가 아닌 단순 실험으로 취급하는 조직일수록 리스크가 커진다. 이 접근 방식은 기술 부족이 아니라 운영 성숙도의 한계로 인해 확장 단계에서 실패하게 된다. MIT 조사에 따르면 생성형 AI 파일럿 프로젝트의 95%는 측정 가능한 비즈니스 성과를 내지 못했으며, 이는 적절한 관리 체계 없이 기존 프로세스에 억지로 적용된 경우가 많기 때문이다. 현장에서 만난 여러 IT 리더들도 동일한 문제를 지적한다. 데이터 분석이나 고객 서비스에 에이전트를 실험적으로 도입한 뒤 문제가 발생하면, 가장 먼저 부딪히는 장애물은 대응을 누가 주도할지 결정하는 일이다. 이러한 혼란은 에이전트의 본질에 대한 오해에서 비롯된다. 기
- Anthropic overtakes OpenAI in enterprise artificial intelligence race
New data reveals Anthropic has surpassed OpenAI in enterprise AI spending, with 34.4% of surveyed companies paying for its products in April. This shift is largely driven by demand for Anthropic's AI coding assistant, Claude Code, marking a significant disruption in the market.
- [Graphic News] ChatGPT leads Korea AI app market
New data from Wiseapp Retail shows ChatGPT has pulled far ahead of rivals in both user base and growth in Korea’s generative AI app market. ChatGPT added 12.37 million monthly active users in March from a year earlier, the largest increase among major apps. Google’s Gemini followed with 7.18 million new users, while Grok AI, Claude and Perplexity posted gains of 1.58 million, 1.34 million and 700,000, respectively. ChatGPT also led in total users, reaching 23.29 million monthly active users as o
- Visiblie
Brand monitoring for AI search results
- TestMu AI Introduces Test.md, an Agent-Native Test Framework for Kane CLI
TestMu AI Introduces Test.md, an Agent-Native Test Framework for Kane CLI Toronto Star