Amazon Q flaw let booby-trapped Git repos execute code, swipe cloud creds
Researchers warn many AI coding assistants now execute commands from project configurations
Latest AI/ML News
33738 matching items
Paystack Launches AI Agent Checkout ‘Index’ in Nigeria, Letting Users Pay Through Claude, ChatGPT and OpenClaw
Paystack, the payments technology company owned by The Stack Group, has launched an experimental product that allows users in Nigeria to check out with supported Paystack merchants using AI agents. Paystack Index, developed with product support from TSG Labs — the group’s venture studio focused on building products using emerging technologies — builds on existing [...]
World Bank and Gates Foundation Launch Initiative to Bring Chichewa Into AI Models in Malawi
Malawi is poised for a major step into the artificial intelligence era after the World Bank, with support from the Gates Foundation, launched an initiative to develop AI models capable of understanding and processing Chichewa — a move expected to transform access to information and public services for millions of people. The project aims to [...]
heise+ | Urlaub ohne Stress: So plant KI Ihre Touren in Minuten – ein Prompt-Leitfaden
Chatbots können die Planung von Ausflügen im Urlaub vereinfachen und beschleunigen. Mit unserer Prompt-Schablone gelingt das schnell und unkompliziert.
Ethiopian Electric Utility Deploys AI to Modernise Power Grid and Reduce Outages
The Ethiopian Electric Utility is carrying out a comprehensive modernization of the country’s power system and is deploying artificial intelligence to create a more reliable and sustainable electricity grid, according to the company’s CEO, Getu Geremew. Geremew, speaking to ENA, said the company is implementing targeted measures to maximize operational efficiency and the transmission capacity [...]
The Case for Model Forensics
If we had a misalignment warning shot, would we be able to tell? Suppose an AI company catches their model taking an egregious action, like deleting oversight code that monitors its actions. Should they sound the alarm? A key piece of evidence to determine what to do next – such as what mitigations to take – is to understand why the model took the action. If the model was just confused (e.g. it may have been trying to reduce latency), a simple mitigation like a regex classifier that blocks destructive actions until a user approves should suffice to prevent the behavior. But if this was intentional subversion, the model will circumvent the regex, and more robust, expensive mitigations are needed. This motivates the need for a follow-up investigation into the concerning behavior, a problem we term model forensics. We recently released a paper that aims to take a concrete step in developing the growing field of model forensics; this post lays out the general case. Motivation If we build AI systems that knowingly cause harm against the developer’s intent, it is critical we recognize this as soon as possible. One plausible way we may do this is through catching bad actions. However, a bad action on its own is not sufficient to conclude misalignment: the model may have done it for benign reasons. This is not just a theoretical concern – in the literature, it is largely the case that when concerning behavior has been dug into, benign explanations have been surfaced. To resolve this…
Xcode 26.6: Google Gemini zieht als Programmierassistent in Apples IDE ein
Mit Xcode 26.6 bringt Apple Google Gemini direkt in die Entwicklungsumgebung. Die Integration war bereits in Xcode 27 beta aufgetaucht, das noch weiter geht.
ITU Chief Praises Mozambique’s Digital Transformation Progress, Including Emerging National AI Strategy
The Secretary-General of the International Telecommunication Union, Doreen Bogdan-Martin, has praised Mozambique’s progress in advancing its national digital transformation agenda, emphasizing the role of connectivity, digitalization and innovation in driving inclusive economic growth. She made the remarks following high-level engagements with Mozambican leadership, including discussions with President Daniel Chapo, where they explored priorities under the [...]
Nigeria Ranks Best in Africa for AI Governance, Government Says at UNESCO Readiness Report Launch
The Federal Government of Nigeria says the country currently ranks best in Africa for artificial intelligence governance, citing a significant climb up the Government AI Readiness Index over the past three years. Minister of Communications, Innovation and Digital Economy Bosun Tijani disclosed the ranking at the launch of the Nigeria Artificial Intelligence Readiness Assessment Methodology [...]
heise+ | KI-Überwachung in der Straßenbahn: Mehr Sicherheit im Fahrgastraum?
Aggression frühzeitig erkennen: Die Bremer Straßenbahn AG testet ein KI-Überwachungssystem. Ein Videobeitrag zu „AI Watch“.
Big Tech is spending trillions on AI. Investors now want proof it will pay off.
"The current push for AI adoption that we're seeing is directly coming from the financial incentives of AI firms," she added. Because of the massive capital expenditures, the hyperscalers and other AI firms are making a "deliberate push for AI everywhere — no matter whether the demand is there or if customers want it or not." The post Big Tech is spending trillions on AI. Investors now want proof it will pay off. appeared first on AI Now Institute .
I have a 100 per cent chance of getting cancer due to a rare gene
A rare variant of a gene called TP53 means Tracy Hutchinson has an extreme risk of developing cancer anywhere in her body, causing endless anxiety and requiring regular whole-body MRIs and other screening
Fine-tuning Language Models on Apple Silicon with MLX
Fine-tune open language models locally on your Mac using MLX. No cloud GPUs or costs required.
Google setzt Verlage bei KI-Lizenzen offenbar unter Druck
Google soll ein Google-News-Programm für Verlage an KI-Rechte für Inhalte knüpfen und das mögliche Auslaufen älterer Vereinbarungen als Druckmittel nutzen.
Water Cooler Small Talk, Ep. 11: Overfitting in RAG evaluation
Why memorizing for the exam doesn't mean you understand the subject The post Water Cooler Small Talk, Ep. 11: Overfitting in RAG evaluation appeared first on Towards Data Science .
Stark bags €500M, Tech.eu Funding Explorer launched, and Luxembourg's big ambitions
This week, we tracked more than 75 tech funding deals worth over €2.1 billion and over 5 exits, M&A transactions, rumours, and related news stories across Europe.If email is more your thing, you c...
From Prototype to Investable Company: What African AI Founders Must Prove
By Sir Roger Jantio, Senior Managing Director & CEO, Sterling Merchant Finance Ltd, Washington, DC For any African AI founder, or would-be founder, this article is a must-read. Not because every founder will agree with every point. Not because there is a single path to building an AI company. But because Africa’s AI moment has [...]
Build interactive PDF text extraction from Amazon S3
In this post, you’ll build a server that extracts text from PDF files in Amazon S3 in real time. This protocol-based approach provides programmatic document access. You’ll walk through the architecture, set up the server, and run interactive document queries. Along the way, you’ll compare this approach with Amazon Textract so you can decide which tool fits your workload.
Databricks will ETL zwischen Datenbanken und Analytics überflüssig machen
LTAP heißt die neue Architektur von Databricks, mit der operative Datenbanken und analytische Systeme nur noch eine gemeinsame Datenbasis benötigen sollen.
pgEdge joins rush to merge OLTP and OLAP storage to support AI
For years, enterprises have maintained separate systems for processing transactional (OLTP) and analytical (OLAP) data, even if that meant moving data between them. However, the rise of autonomous agents and AI applications needing immediate access to data while generating volumes of operational data themselves, has exposed the cost and complexity of maintaining those separate systems. The industry’s response has been quick, with data warehouse and database vendors proposing a wave of competing approaches to collapsing those data silos. In the past few weeks Databricks unveiled LTAP and EDB introduced converged analytics , while late last year Snowflake launched pg_lake , all of which offer different blueprints for bringing transactional, analytical and AI workloads closer together. Now it’s the turn of distributed PostgreSQL provider pgEdge, which has introduced a beta version of ColdFront , a PostgreSQL-native hot-and-cold data tiering architecture that automatically moves older data into Apache Iceberg object storage while keeping PostgreSQL as the only database that applications need to interact with. In ColdFront’s architecture, hot and cold refer to newer and older data, respectively. The approach of keeping PostgreSQL as the primary interface is what sets ColdFront apart from the other architectures emerging in this space, differing in where the center of gravity for data lies, according to analysts. Databricks’ LTAP keeps operational applications connected to a lakeh…
How Cara pioneers domain-specific AI for enterprise insurance brokerages with AWS
In this post, we explore how Cara, built in cooperation with AWS, addresses these challenges. We walk through the technical design decisions and the AWS services that support the solution. We also share measurable outcomes Cara has delivered for enterprise brokerages.
Oracle promises to open up MySQL governance, but the community wants guarantees
Open source advocates remain concerned over lack of binding commitments
Production-grade AI agents for financial compliance: Lessons from Stripe
In this post, you learn how Stripe built a production-grade AI agent system for financial compliance. We cover the technical architecture of Stripe’s ReAct agent framework and the infrastructure decisions behind a dedicated agent service. We also discuss the role of human oversight in maintaining accountability, and key lessons about task decomposition, orchestration patterns, and cost optimization through prompt caching. By the end, you will understand how to design agentic systems that scale compliance operations without compromising quality or auditability.
pgEdge joins rush to merge OLTP and OLAP storage to support AI
For years, enterprises have maintained separate systems for processing transactional (OLTP) and analytical (OLAP) data, even if that meant moving data between them. However, the rise of autonomous agents and AI applications needing immediate access to data while generating volumes of operational data themselves, has exposed the cost and complexity of maintaining those separate systems. The industry’s response has been quick, with data warehouse and database vendors proposing a wave of competing approaches to collapsing those data silos. In the past few weeks Databricks unveiled LTAP and EDB introduced converged analytics , while late last year Snowflake launched pg_lake , all of which offer different blueprints for bringing transactional, analytical and AI workloads closer together. Now it’s the turn of distributed PostgreSQL provider pgEdge, which has introduced a beta version of ColdFront , a PostgreSQL-native hot-and-cold data tiering architecture that automatically moves older data into Apache Iceberg object storage while keeping PostgreSQL as the only database that applications need to interact with. In ColdFront’s architecture, hot and cold refer to newer and older data, respectively. The approach of keeping PostgreSQL as the primary interface is what sets ColdFront apart from the other architectures emerging in this space, differing in where the center of gravity for data lies, according to analysts. Databricks’ LTAP keeps operational applications connected to a lakeh…
A ‘casually horrifying, unexpectedly compassionate’ vision of British decline
M. John Harrison has written a dystopian sci-fi novel in which globalization is only a memory.
Schöne neue Suchwelt: Warum Google und Co. bald für KI-Fehler haften sollen
Wer fragt, kriegt fertige KI-Antworten statt Linklisten. Laut einem Rechtsgutachten für die Landesmedienanstalten sollen die Tech-Riesen dafür voll einstehen.
Anthropic’s Mythos mess is only getting worse
It's been two weeks since Anthropic took its Mythos-class models offline after a Friday evening ultimatum from the Trump administration. The company sprang into action immediately, sending a barrage of executives to Washington, DC. But updates have been suspiciously lacking, with no resolution in sight. Anthropic declined to comment multiple times this week about the […]
OpenAI staggers AI model release after Trump administration request
Sam Altman announces limited preview of GPT 5.6 in move that echoes launch of Anthropic’s Mythos Business live – latest updates OpenAI is staggering the release of its latest AI model after a request from the US government, in a move echoing the launch of Anthropic’s Mythos product. The company behind ChatGPT signalled its dissatisfaction with the move, saying that doing so keeps the best AI tools from “users, developers, enterprises, cyber defenders, and global partners who need them”. Continue reading...
heise+ | Fable 5 Notaus: So zog die US-Regierung die Anthropic-KI aus dem Verkehr
Knappe vier Tage nach Veröffentlichung nahm Anthropic sein KI-Modell vom Netz. Der Fall zeigt, wie schnell die US-Regierung den Notausschalter betätigen kann.
The AI Agent Tech Stack Explained
• <a href="https://www.
Paging Charity! How can engineering leaders avoid becoming Bond villains?…
If you want your values to spread throughout the industry, the best thing you can possibly do is succeed and make others want to imitate you.
US reportedly wants OpenAI to delay GPT-5.6 rollout after restrictions on Anthropic’s models
While the White House asked OpenAI to limit the GPT 5.6 release, the company revealed plans to make it available only to a small group of partners with the govt approving access "customer by customer." The post US reportedly wants OpenAI to delay GPT-5.6 rollout after restrictions on Anthropic’s models appeared first on MEDIANAMA .
One man, two kernels, and a lot of RISC-V
A homebrew PC and mini-mainframe were only the warm-up for Yuri Zaporozhets' latest operating system
‘Digital ID cards’: China moves to regulate AI agents with unified identity system
China is establishing an identity system for artificial intelligence agents, as part of new national standards released on Friday to regulate the next frontier of autonomous technology. The State Administration for Market Regulation (SAMR) unveiled the standard for “Artificial Intelligence Agent Interconnection”, aiming to establish a “closed-loop system” with a unified identity management framework for all AI agents, according to a report from state broadcaster China Central Television...
How to fix model parameter estimates MPlus bi-factor CFA
If not clear from the title, this code is in MPlus. When I run the following code, I get the following issue, but I'm not sure how to fix it. The output doesn't give me fit statistics and I'm not sure where or why the parameters don't work, as for regular CFA and ESEM they did, so it's not a direct data issue. Thanks in advance! TITLE: Bi-CFA for EDE-Q; DATA: FILE IS EDE-Q.csv; VARIABLE: NAMES = Q1-Q12, Q19-Q28; ! These are the variables that contribute to the EDE-Q, labelling each coloumn in the shortened data set MISSING = *; ! Giving blanks a value so MPlus can see where they are for full maximum likelihood indicator, automatically fills in any missing data gaps CATEGORICAL ARE Q1-Q12, Q19-Q28; ! Specifies that data is categorical, not continuous MODEL: GeneralFactor BY Q1-Q12, Q19-Q28; ! General factor measured by all items F1 BY Q1-Q5; ! Each item on it's predetermined sub-scale, Restraint, residual variance being measured in addition to general factor F2 BY Q7, Q9, Q19-Q21; ! Eating Concern F3 BY Q6, Q8, Q10, Q11, Q23, Q26-Q28; ! Shape Concern F4 BY Q8, Q12, Q22-Q25; ! Weight Concern GeneralFactor WITH F1-F4@0; ! General and specific factors are not correlated and so set to 0, sub-scales are residual factors F1-F4 WITH F1-F4@0; ! Typically specific factors shouldn't be correlating OUTPUT: SAMPSTAT, STDYX; ! Sample statistics and standardised outputs THE STANDARD ERRORS OF THE MODEL PARAMETER ESTIMATES COULD NOT BE COMPUTED. THE MODEL MAY NOT BE IDENTIFIED. CHECK YOUR M…
Nigerian lawmakers approve police overhaul
The decision marks a new effort to tackle an endemic, decades-old security crisis.
Uber India & South Asia President Prabhjeet Singh steps down, likely to join OpenAI as India MD
Uber India and South Asia President Prabhjeet Singh has stepped down after spending nearly 11 years with the ride hailing company. Confirming the development, an Uber spokesperson said, "India is one of Uber's most important markets globally, an important driver of innovation and long term growth. The strength of our business today reflects the incredible team and foundation built over the years. We thank Prabhjeet for his leadership and lasting contributions in his decade long journey with Uber. We remain deeply committed to our next phase of growth in India." Media reports suggest that Singh is set to join OpenAI as India MD. Singh joined Uber in August 2015 after a stint at McKinsey & Company, where he was an Associate Partner. Over the years, he held multiple leadership roles before being appointed President of India and South Asia in July 2020. As President, Singh led Uber's mobility operations across India, Bangladesh and Sri Lanka. Under his leadership, the company diversified beyond four wheeler ride hailing by expanding offerings such as Auto, Moto and Shuttle, while strengthening collaborations with public transport authorities and digital public infrastructure initiatives. Singh's departure comes days after Uber CEO Dara Khosrowshahi visited India and announced the company's first data centre in the country in partnership with Adani Group, signalling a deeper investment in its technology and infrastructure footprint. In February, Uber also infused nearly Rs 3,000…
5 Agentic Workflows to Automate Your Data Science Pipeline
This article covers five concrete agentic workflows, one for each major stage of a data science pipeline.
Amplify the Expert: A Philosophy for Building Enterprise RAG
Enterprise Document Intelligence [Vol.1 #M1] - The thesis behind every architectural choice in this series The post Amplify the Expert: A Philosophy for Building Enterprise RAG appeared first on Towards Data Science .
South Africa corruption investigation widens
Business Vusimuzi Matlala offered to provide evidence against members of the goverment underscoring the breadth of a huge graft probe in the continent’s biggest economy.
US to provide new experimental Ebola drug
Slowing the virus’ spread has been hampered by the fact that there is no known treatment or vaccine for the new Bundibugyo strain.
How to Fine-Tune RF-DETR Keypoints on Custom Data
A step-by-step guide to fine-tuning RF-DETR Keypoint on a custom basketball court dataset, from COCO pretrained inference to training, evaluation, and broadcast-video inference.
Notion kills its Gmail client after AI agents keep humans from troubling inbox
More than half of users now let bots handle email, so service is headed for shutdown
Fakes im Netz: So entlarvt jeder KI-generierte Videos (von uns getestet)
KI-Fakes überschwemmen Social Media. c't 3003 hat sich selbst getestet und liefert die 6-Punkte-Checkliste, mit der jeder Deepfakes zuverlässig entlarvt.
KI-Update Deep-Dive: Synthetische Bilder gegen den Bias
KI-Modelle zur Hautkrebserkennung sind oft voreingenommen. Das Projekt "MedGenAI" des DFKI will das Problem mit gezielt erzeugten synthetischen Bildern lösen.
Robotaxis drive miles just to get cleaned and charged; this new startup wants to fix that
Aseon Labs, which came out of Y Combinator's 2026 spring cohort, has raised $10 million from Crane Venture Partners and others.
Why your top ranking no longer means you’re winning
AI and personalization are replacing one-size-fits-all rankings with unique customer journeys, forcing marketers to rethink how they measure visibility. The post Why your top ranking no longer means you’re winning appeared first on MarTech .
China’s transnational interference threatens digital rights globally
China’s transnational interference threatens digital rights globally H.Seidl Fri, 06/26/2026 - 14:56 picture alliance / NurPhoto | Jaap Arriens Comment Jun 26, 2026 4 min read China’s transnational interference threatens digital rights globally Beijing’s coercive use of digital tools and economic leverage undermines international efforts to regulate digital technologies, say Daria Impiombato and Wendy Chang. Signs are mounting that the Chinese government is expanding its transnational repression both in terms of tools and targets. The first half of 2026 has seen evidence of online and offline attempts to silence overseas critics that cross its political red lines. Only in May, an AI-generated harassment campaign against Europe-based human rights researcher Laura Harth, known for her work exposing China’s overseas police stations, was made public. The campaign, which relied on misogynistic and sexualized images, shows how Beijing is incorporating generative AI into its transnational repression efforts, allowing new forms of scalable, personalized attacks aimed at damaging the reputation of critics abroad. But attempts to silence individuals have also widened to target global civil society collectively. Another recent victim of a reported Chinese government campaign was an entire conference dedicated to advancing digital rights for all – the rights people should enjoy online, including privacy, freedom of expression, access to information and protection from unlawful surveilla…
SAP aligns commerce data for AI personalisation
SAP aligns fragmented commerce data structures to enable operational AI personalisation at the execution layer. Enterprise leadership routinely establishes objectives to anticipate customer requirements and deliver relevant interactions across digital touchpoints. However, the actual infrastructure running inside these enterprises fails to support systematic execution at the required volume. Recommendation engines display generic product listings because […] The post SAP aligns commerce data for AI personalisation appeared first on AI News .
Ancient human DNA found on cave art for the first time
DNA from ancient humans has been found on a prehistoric cave painting and on cave walls, demonstrating the potential to one day identify individual artists and resolve the debate over Neanderthals' artistic abilities