AI/ML News & Innovations Hub

AI/ML news, top picks, and generated innovation digests.

★ Visit ai-karthik.com
426Sources
90413News Items
8Top Picks
445Blogs
runningLast Run

Latest AI/ML News

90413 matching items

CIO AI 2026-11-10 10:00 UTC Score 49.0 USR-0125-20261110-global-ai-ne-b9f7b514 Full article

Culture is the new AI barrier: Why trust and adoption are more than technical challenges

There’s intense focus among business leaders on building the right IT foundation to support AI across the organization. There’s a rush to provision computing resources, unify data sources and modernize security, among other investments. But the implementation of AI is both a technical and organizational challenge. To actually get value, employees must know how to effectively wield the tools and understand their role in the emerging agentic processes, both as a security check and a mentor. Businesses need to trust they have the right infrastructure in place to scale, and employees have to trust the outputs are accurate — particularly as AI agents operate with growing independence. Increasingly, AI can move beyond simply monitoring systems. Agents can start to act like a site reliability engineer, for example, retrying failed jobs, identifying problems and initiating self-healing actions. Or in post-trade processing, AI agents can automate exception management, a traditionally manual process that requires information to move back and forth between financial institutions and counterparties. To manage both the cultural and technical changes required, companies should think of AI adoption in three layers: foundations, trust and adoption. Foundations AI success starts with the fundamentals. At the most basic level, it’s about making sure the resources are there to power the AI agents. But soon enough, any company will be able to buy or rent enough hardware to run AI at scale. The…

iAfrica 2026-11-10 09:59 UTC Score 25.0 AI-151-20261110-regional-ai--5aa73947 Full article

South African Author Says She Has Published 370 AI-Assisted Novels and Earned $1m This Year

A South African writer using the pen name Coral Hart says she has published more than 370 novels with the help of AI since February 2025, selling over a million copies and earning more than $1 million this year. The figures are her own and have not been independently verified. Hart told AFP she turned [...]

iAfrica 2026-11-10 09:55 UTC Score 43.0 AI-151-20261110-regional-ai--c97e10f5 Full article

AWS Commits $100m in Cloud and AI Credits to Health Projects, With Kenyan Providers Among Participants

Amazon Web Services is committing $100 million in AI and cloud resources over five years to help healthcare organisations reach underserved communities, with several Kenyan providers among those already using its tools. The AWS Health Impact Initiative will provide cloud credits, AI tools and agentic services to health nonprofits, government agencies, research institutions and technology [...]

iAfrica 2026-11-10 09:52 UTC Score 33.0 AI-151-20261110-regional-ai--5a84bbcb Full article

AI And Blockchain Are Coming To African Farms

The use of digital agriculture is expanding across Africa. Tools like artificial intelligence (AI), sensors and blockchain are helping farmers receive information about weather, pests, markets and farm management. AI can analyse images and other farm data to detect problems or support decisions. Digital traceability systems can record where food comes from and what happens [...]

iAfrica 2026-11-10 09:48 UTC Score 25.0 AI-151-20261110-regional-ai--d5f22c15 Full article

Meta Expands AI Age Detection on Instagram Globally, Placing Millions More Teens in Restricted Accounts

Meta is expanding AI-powered age detection on Instagram to users worldwide, a move it says will place millions more teenagers into its restricted Teen Accounts. The system disregards the date of birth users enter and instead uses images to estimate their actual age. Users identified as teenagers are moved into Teen Accounts, which carry additional [...]

Thailand moves to block Bitget from December 5 over alleged unlicensed crypto trading
Medianama AI 2026-11-10 09:44 UTC Score 40.0 USR-0211-20261110-regional-new-11350b3a Full article

Thailand moves to block Bitget from December 5 over alleged unlicensed crypto trading

Thailand's SEC has filed a criminal complaint against Bitget's operator for allegedly running an unlicensed crypto exchange. Authorities plan to block access from December 5, 2026. The post Thailand moves to block Bitget from December 5 over alleged unlicensed crypto trading appeared first on MEDIANAMA .

LessWrong AI 2026-11-10 09:40 UTC Score 76.0 USR-0152-20261110-community-fo-633d13bb

Honeypots for AI Agent Insider Threats

TLDR: We propose a new AI control method: Honeypots - fake opportunities for malicious actions - placed inside and around an agent during production usage. Previous work uses honeypots in evaluations. We argue they should run continuously during deployment. Such honeypots could detect misbehavior, incriminate malicious agents, and deter them from attempting attacks. We lay out a research agenda for making this happen. This post is adapted from our IAPS report of the same name. Link to the full paper: https://www.iaps.ai/research/honeypots-for-ai-agent-insider-threats Why agent honeypots? Agents deployed inside organizations get broad access to do their work, but can also misuse these affordances. If agents make mistakes, are compromised by attackers, or follow misaligned objectives, they effectively become insider threats. We propose that honeypots should be used to defend against such insider threats. While previous research uses honeypots as propensity evaluations, we argue that frontier AI companies should place fake opportunities for malicious actions in and around their agents’ environments while conducting inference during RL, evals, and production usage. We propose three functions that honeypots could serve: Detection : Honeypots can act as tripwires, alerting security staff when an agent takes unauthorized action. Example: Introduce apparent misconfigurations in sandboxes . Any agent that attempts to exploit them is immediately paused. Incrimination : Honeypots can e…

InfoWorld AI 2026-11-10 09:29 UTC Score 59.0 USR-0126-20261110-global-ai-ne-8cc69097 Full article

Anthropic unveils Claude Fable 5.2, Mythos 5.2 with coding gains and lower prices

Anthropic has introduced Claude Fable 5.2 and Mythos 5.2, claiming significant improvements in handling complex software engineering tasks while cutting API prices by 20% compared with the previous versions. Fable 5.2 and Mythos 5.2 share the same underlying model but have different safeguards. Fable is broadly available, while direct access to Mythos is restricted to vetted cybersecurity and life sciences organizations to limit potential misuse. The models are both priced at $8 per million input tokens and $40 per million output tokens, down from $10 and $50, respectively. Anthropic said the newer versions can generate output up to 30% faster than their predecessors. The company reported substantial gains in coding performance. Fable 5.2 scored 49.1% on the FrontierCode 2.0 benchmark at its Xhigh effort setting, compared with 32.4% for Fable 5.1. The benchmark assesses whether maintainers of open-source projects would accept AI-generated changes to their code. Performance and costs could entice enterprises The models’ improved performance and lower prices could make long-running AI agents viable for software migrations and scientific research, according to Lian Jye Su , chief analyst at Omdia. Higher task completion rates could further reduce deployment costs, he said. Anthropic said Fable 5.2 at its “High effort” setting achieved a coding benchmark score comparable to Opus 5.5 at its highest setting, at roughly half the cost per task. The comparison suggests that Fable cou…

Meghan Markle makes first public appearance since father's arrest
Euronews AI 2026-11-10 09:25 UTC Score 37.0 AI-164-20261110-regional-ai--fc8b42a5 Full article

Meghan Markle makes first public appearance since father's arrest

The Duke and Duchess of Sussex have arrived in Brussels for an event highlighting the dangers of online harm to children. Meghan's appearance marks her first public engagement since her father, Thomas Markle Sr., was charged with possessing child sexual abuse material in the US.

챗GPT, 주요 AI 챗봇 이용량 81% 차지…클로드 급성장에도 독주
Korea AI Times 2026-11-10 09:14 UTC Score 40.0 USR-0048-20261110-global-ai-ne-0977d26c Full article

챗GPT, 주요 AI 챗봇 이용량 81% 차지…클로드 급성장에도 독주

구글 \'제미나이\'와 앤트로픽 \'클로드\'의 추격에도 오픈AI의 \'챗GPT\'가 글로벌 AI 챗봇 시장에서 압도적인 지위를 유지하고 있는 것으로 나타났다. 경쟁 서비스들이 빠르게 성장하는 가운데 챗GPT도 올해 상반기 하락했던 점유율을 최근 다시 회복하면서 주요 AI 챗봇 이용량의 81% 이상을 차지했다.시장조사 업체 그래파이트(Graphite)는 9일(현지시간) 시밀러웹 데이터를 분석한 보고서를 통해 챗GPT가 전 세계 주요 AI 챗봇 6종의 웹·앱 이용 세션 가운데 81.2%를 차지했다고 밝혔다.조사 대상에는 챗GPT와 클로드, 제미

클로드 페이블 5.2, 전자기기 3종 설계…회로 수정 없이 시제품 제작
Korea AI Times 2026-11-10 09:05 UTC Score 40.0 USR-0048-20261110-global-ai-ne-310e4a27 Full article

클로드 페이블 5.2, 전자기기 3종 설계…회로 수정 없이 시제품 제작

앤트로픽이 \'클로드\'로 전자기기 3종을 설계하고, 회로 수정 없이 모두 작동하는 시제품을 제작하는 데 성공했다. 오픈AI가 최근 AI를 활용한 전자기기 제작 사례를 공개한 데 이어, AI가 실제 제조 조건까지 고려해 하드웨어 개발을 수행하는 능력을 보여줬다.앤트로픽은 9일(현지시간) 공식 기술 보고서를 통해 최신 모델 \'클로드 페이블 5.2\'가 전자회로 설계부터 부품 배치, 펌웨어 작성, 제조 가능성 검토까지 하드웨어 개발의 주요 과정을 수행했다고 밝혔다. 이를 기반으로 실제 작동하는 전자기기 시제품 3종을 제작했다.앤트로픽은 전자

InfoWorld AI 2026-11-10 09:00 UTC Score 69.0 USR-0126-20261110-global-ai-ne-8e4079a1 Full article

Anthropic and OpenAI are coming for the development workflow

Running a gold mine can be profitable. Finding and digging the next one is expensive. As analyst Benedict Evans argues , Anthropic’s business has the same challenge, with more than 80% of its third-quarter revenue coming from models released in the preceding six months. A lab can make good money serving its best model while spending heavily on the one it will need next, but it could make amazing money if it could go 18 months between major training runs instead of six. Reality, however, bites. This pressure to maximize profitability helps explain why frontier labs like Anthropic and OpenAI are aggressively adding functionality to own more of the developer’s workflow. On November 4, OpenAI announced an agreement to buy Pierre , the company behind Code Storage and the open source libraries Codex already uses to browse files and review changes. OpenAI wants Codex to be where developers work, from the first branch an agent creates through deployment of the application. There’s money in extending the life of that gold mine. The frontier labs are betting that better models will win a developer’s business and that they can keep that business by taking over the complete development workflow. But developers should be wary of turning a preference for this quarter’s best model into a much longer commitment to the company supplying it. As I’ve noted , we need places like GitHub where competing agents can work together, and a practical way to change our minds about which ones we use. The…

InfoWorld AI 2026-11-10 09:00 UTC Score 34.0 USR-0126-20261110-global-ai-ne-e34917f7 Full article

Panic over cloud capacity is out of hand

I read the recent piece about AWS cloud capacity being “largely sold out” through 2027 , and my first reaction was fairly simple: Well, that’s that. I guess I’ll stop recommending Amazon Web Services to any client looking to host AI workloads. Clearly, the cloud is full. Please take a number. Your GPU will be available sometime after the next presidential election, assuming the data center hamsters are still running. Of course, I suspect I can still log in to the AWS console today and provision some kind of compute for AI workloads. Maybe not a massive, multiyear, hyperscale GPU allocation suitable for training the next frontier model. Maybe not the exact reserved-capacity block a large AI lab wants to lock down through 2028. But that is not the same thing as saying the cloud provider is “sold out.” The AI infrastructure conversation has gone off the rails. Demand is high. Supply is constrained. GPUs are expensive. Power is a problem. Data centers do not magically appear because somebody wrote “AI strategy” on a whiteboard. All of that is true. But “capacity is tight” and “capacity is sold out” are not the same sentence. The confused market Many cloud providers and infrastructure companies seem very eager to tell the market that their AI capacity is committed for years. I understand the motivation. It signals demand. It reassures investors. It tells customers, partners, and Wall Street that the AI boom is real and that everyone wants what they are selling. It is the infrastr…

CIO AI 2026-11-10 09:00 UTC Score 53.0 USR-0125-20261110-global-ai-ne-dca91199 Full article

What kind of management style is needed in the age of AI?

As AI increasingly takes over routine tasks and basic information processing, simply using the same AI tools as competitors is no longer enough to differentiate a company. Competitive advantage is increasingly rooted in what makes each organization unique: its philosophy, culture, and accumulated tacit knowledge. For business leaders, including CIOs, the central question is shifting from “How can AI improve efficiency?” to “How can we foster continuous growth for both our people and our organizations in the AI era?” Two influential perspectives offer valuable insights into this challenge. The first is Founder Mode, an essay published in September 2024 by Paul Graham , co-founder of Y Combinator , the influential Silicon Valley startup accelerator. The second is The Founder’s Mentality, a management framework developed by Bain & Company through its extensive research into the drivers of sustained corporate growth and the crises that can undermine it. Both address a fundamental challenge: as organizations expand, how can they preserve the sense of ownership, entrepreneurial energy, and intuitive decision-making that defined their early success? In this article, I explore how Founder Mode can serve as a practical leadership approach for sustaining the principles of the Founder’s Mentality —and why this combination may become increasingly important for organizational growth in the AI era. Founder Mode : Rethinking the conventional wisdom of delegation As an organization grows, l…

보스턴 다이내믹스, 바퀴 달린 '아틀라스' 공개…360도 회전 점프까지
Korea AI Times 2026-11-10 08:57 UTC Score 40.0 USR-0048-20261110-global-ai-ne-f34c7760 Full article

보스턴 다이내믹스, 바퀴 달린 '아틀라스' 공개…360도 회전 점프까지

보스턴 다이내믹스가 휴머노이드 로봇 \'아틀라스(Atlas)\'의 발을 바퀴로 교체해 빠르게 주행하고 360도 회전 점프까지 하는 모습을 공개했다. 강화학습(RL) 기반 제어 기술을 적용해 새로운 이동 방식을 익히도록 한 것으로, 로봇의 구동 장치와 AI 제어 모델의 성능을 검증하기 위한 연구용 실험이다.보스턴 다이내믹스는 9일(현지시간) 공식 영상을 통해 바퀴를 장착한 아틀라스가 실제 실험 공간에서 빠르게 주행하고 방향을 전환하는 모습을 선보였다. 측면 질주와 360도 회전 점프는 물론, 한쪽 바퀴로 균형을 유지하거나 바닥에 누운 상

텐센트도 'AI 자기 개선' 연구 본격화…미국 이어 RSI 경쟁 가세
Korea AI Times 2026-11-10 08:50 UTC Score 40.0 USR-0048-20261110-global-ai-ne-66861c3f Full article

텐센트도 'AI 자기 개선' 연구 본격화…미국 이어 RSI 경쟁 가세

오픈AI와 앤트로픽 등 미국 주요 AI 기업들이 재귀적 자기 개선(RSI)을 차세대 AI 개발의 핵심 과제로 제시하는 가운데, 중국 텐센트도 AI가 자신의 작업 결과를 활용해 성능을 높이는 기술을 선보였다. 지난달 공개된 연구가 최근 개발자 커뮤니티에서 다시 주목받으면서 중국도 AI의 자기 개선 기술 경쟁에 뛰어들었다는 평가가 나온다.텐센트의 하이 프론티어(Hy Frontier) 연구팀과 싱가포르국립대학교(NUS) 등 공동 연구진은 지난달 15일 아카이브에 \'터미널-RSI(Terminal-RSI)\' 논문을 발표했다. 지난달 30일

스페이스X, 텍사스에 32조 AI 데이터센터·가스발전소 건설 추진
Korea AI Times 2026-11-10 08:48 UTC Score 40.0 USR-0048-20261110-global-ai-ne-ec958939 Full article

스페이스X, 텍사스에 32조 AI 데이터센터·가스발전소 건설 추진

스페이스X가 미국 텍사스주 스타베이스에 240억달러(약 32조원) 규모의 AI 데이터센터와 가스발전소 건설을 추진한다. 자체 전력 공급 시설까지 확보해 AI 연산 수요에 대응하고 데이터센터 가동에 필요한 전력망 연결 대기 시간을 줄이려는 전략으로 풀이된다.9일(현지시간) 로이터에 따르면, 텍사스주 스타베이스시는 최근 공개한 공청회 공고를 통해 스페이스X가 가스발전소와 슈퍼컴퓨팅 시설 건설을 추진한다고 밝혔다.사업 부지는 스페이스X의 로켓 발사기지가 있는 스타베이스 내 444에이커(약 180만m²) 규모다. 초기 단계에 예상되는 자본

KT, 3분기 실적 주춤...B2B·AX로 반등 노려
Korea AI Times 2026-11-10 08:41 UTC Score 35.0 USR-0048-20261110-global-ai-ne-bb6f78ac Full article

KT, 3분기 실적 주춤...B2B·AX로 반등 노려

KT가 매출 및 영업이익 동반 감소라는 3분기 성적표를 받아 들었다.KT는 올해 3분기 연결 기준 매출 6조9975억 원, 영업이익 5053억 원을 기록했다고 10일 밝혔다. 지난해 같은 기간 대비 각각 1.8%, 6.1% 줄었다.이 같은 실적 하락에 대해 전년 3분기 강북본부 부지 개발에 따른 일회성 분양이익이 반영된 기저효과 때문이라고 KT 측은 설명했다. 별도 기준 영업이익은 2993억원으로, 7월 종료된 고객 보답 프로그램과 AX 사업 비용 증가가 영향을 미쳤다.다만, 본업은 회복 신호를 보이고 있다고 밝혔다. 무선 서비스

클로드, 밤새 로봇 팔로 수채화 완성…물감도 직접 섞었다
Korea AI Times 2026-11-10 08:40 UTC Score 38.0 USR-0048-20261110-global-ai-ne-407bbaab Full article

클로드, 밤새 로봇 팔로 수채화 완성…물감도 직접 섞었다

앤트로픽의 \'클로드\'가 로봇 팔을 이용해 밤새 수채화를 완성한 사례가 공개됐다. 특히 작업 도중 필요한 색상이 없다는 사실을 알아차리고 직접 물감을 혼합해 그림을 이어가면서, 범용 AI가 실제 도구를 활용해 물리적 작업을 수행하는 능력을 보여줬다.베를린에서 활동하는 디자인 엔지니어 닐(nil)은 지난 6일(현지시간) X에 클로드가 가을 나무를 그리는 영상을 공개했다. 잠들기 전 로봇 팔에 연결된 클로드에게 그림을 요청했고, 다음 날 아침 실제 수채화 한 점이 완성돼 있었다는 설명이다.작업 도중에는 예상치 못한 문제도 해결했다. 새벽

Stack Overflow AI Blog 2026-11-10 08:40 UTC Score 46.0 USR-0063-20261110-ai-specialis-a4e965ad Full article

Customer support is the gnarliest AI problem

Ryan chats with Ahmad Al-Dahle, CTO at Airbnb, about deploying multi-party AI agents to tackle the uniquely gnarly problem of customer support, how a moving target of rigorous evaluations helps engineers balance latency against accuracy, and the ways in which AI coding tools demand an even higher level of judgment from software engineering teams.

Is Burnham’s No 10 North working out?
Politico Europe AI 2026-11-10 08:35 UTC Score 37.0 AI-170-20261110-regional-ai--a10d8540 Full article

Is Burnham’s No 10 North working out?

Is the battle over welfare reforms about to intensify? With the Milburn Report on youth unemployment just a week away, Sam Coates and Anne McElvoy examine the growing pressure on Labour to tackle the issue. And with the head of the civil service, Antonia Romeo, preparing Whitehall for a spending squeeze, the Conservatives are planning […]

Amit Shah approved Make in India exemption for NIA’s Cellebrite phone-cracking tool
Medianama AI 2026-11-10 08:22 UTC Score 42.0 USR-0211-20261110-regional-new-ae79b15a Full article

Amit Shah approved Make in India exemption for NIA’s Cellebrite phone-cracking tool

India has no published protocol governing forensic phone extraction, even as agencies procure tools capable of accessing extensive smartphone data. The post Amit Shah approved Make in India exemption for NIA’s Cellebrite phone-cracking tool appeared first on MEDIANAMA .

바이트댄스, 100회 수정해도 이미지 유지하는 '시드림 5.5' 공개
Korea AI Times 2026-11-10 08:20 UTC Score 37.0 USR-0048-20261110-global-ai-ne-8747f8ea Full article

바이트댄스, 100회 수정해도 이미지 유지하는 '시드림 5.5' 공개

바이트댄스가 이미지를 최대 100회 연속 수정해도 원래 모습을 유지하고, 투명 배경과 정교한 텍스트까지 구현하는 차세대 이미지 생성 모델을 선보였다. 반복 편집 과정의 안정성과 이미지 사실성을 강화해 전문 디자인과 콘텐츠 제작 현장에서 활용도를 높이는 데 초점을 맞췄다.바이트댄스는 9일 이미지 생성 모델의 편집 안정성과 사실성, 텍스트 표현력을 강화한 신형 모델 \'시드림 5.5(Seedream 5.5)\'를 공개했다.이번 모델의 핵심 개선 사항은 크게 세 가지다. 먼저 여러 차례 편집하는 과정에서 수정하지 않은 영역을 픽셀 단위로 보

BMW, '챗GPT'로 차량 추천부터 시승 예약까지…AI 쇼핑 확대
Korea AI Times 2026-11-10 08:13 UTC Score 40.0 USR-0048-20261110-global-ai-ne-4304959d Full article

BMW, '챗GPT'로 차량 추천부터 시승 예약까지…AI 쇼핑 확대

BMW가 \'챗GPT\'를 통해 고객이 원하는 차량을 추천받고 금융 조건 비교부터 시승 예약까지 진행할 수 있는 자동차 쇼핑 서비스를 선보였다. 기존 차량 검색과 옵션 구성 기능을 실제 판매점 예약 시스템까지 연결해 AI를 활용한 자동차 구매 과정을 확대했다.BMW 북미 법인은 9일(현지시간) 챗GPT에서 신차 검색부터 맞춤형 차량 구성, 금융 조건 확인, 시승 예약까지 한 번에 처리할 수 있는 차량 쇼핑 플러그인을 미국에 출시했다고 발표했다. AI 챗봇 안에서 자동차 구매 상담을 시작해 실제 판매점(딜러십) 방문 예약까지 마칠 수 있

오픈AI, AI 연구용 '킬 스위치' 공개…35초 내 출력 차단
Korea AI Times 2026-11-10 08:02 UTC Score 38.0 USR-0048-20261110-global-ai-ne-4e83b742 Full article

오픈AI, AI 연구용 '킬 스위치' 공개…35초 내 출력 차단

오픈AI가 연구 중인 AI 모델에서 위험한 행동이나 보안 문제가 발견되면 추가 출력을 신속하게 차단하는 중앙 통제 시스템을 공개했다. 모델이나 연구 프로그램 자체를 종료하지 않고 문제가 발생한 작업의 실행 권한만 차단해, 다른 연구에는 영향을 주지 않으면서 안전 문제에 대응할 수 있도록 설계했다.오픈AI는 9일(현지시간) 공식 블로그를 통해 AI 연구 인프라 전반에서 실행되는 작업을 추적하고, 안전이나 보안 문제가 발생하면 해당 작업을 중단하는 \'컨트롤 룸(Control Room)\'을 소개했다. 모델 학습을 위한 에이전트 샘플링부터

[게시판] AWS, 의료 격차 해소에 5년간 1억달러 기술 지원 등 단신
Korea AI Times 2026-11-10 08:00 UTC Score 40.0 USR-0048-20261110-global-ai-ne-2915ba62 Full article

[게시판] AWS, 의료 격차 해소에 5년간 1억달러 기술 지원 등 단신

■ 아마존웹서비스(AWS)는 의료 소외 계층의 의료 격차 해소 솔루션을 구축하는 기관에 AI와 클라우드 인프라를 지원하는 ‘AWS 헬스 임팩트 이니셔티브’를 발표했다고 밝혔다. 향후 5년간 1억달러 규모의 기술을 지원하며 비영리단체와 정부·연구 기관 및 기업 등이 지원할 수 있다. ▲의료 서비스 접근성 확대 ▲여성·아동 건강 연구 ▲건강한 식품 접근성 등 3개 분야를 중점 지원한다.■ LG유플러스(대표 홍범식)는 GS네오텍과 엔드투엔드(E2E) 보안 사업 협력을 위한 전략적 업무협약을 체결했다고 밝혔다. 양사는 LGU+의 네트워크·