Author: Art Ryan

Ajman has moved beyond the familiar government chatbot. The emirate has completed what officials describe as the UAE’s first government transaction powered by agentic AI, using an intelligent assistant to renew a trade licence with minimal manual intervention. Instead of waiting for a business owner to notice an approaching deadline, gather documents and move between different government platforms, the system can identify the upcoming renewal and initiate the process proactively. That distinction matters. This is not simply AI answering questions faster. It is AI taking responsibility for moving a government transaction forward. The AI Agent Contacts Businesses Before Their Licences…

Read More

Sakana AI has released Fugu Ultra v1.1, an updated version of its multi-agent AI system built for difficult coding, research and reasoning tasks. It is not a dramatic rebuild. There is no new architecture to learn, no fresh API migration and, perhaps more importantly, no price increase. Instead, Sakana AI has changed the models working behind the scenes. Fugu Ultra v1.1 now incorporates newer frontier models inside its orchestration system, allowing the platform to improve without forcing developers to rebuild their integrations. That may sound like a routine update. It is not quite that simple. Fugu Ultra does not behave…

Read More

Alibaba’s Qwen team is pushing deeper into synthetic speech with Qwen Audio 3.0 TTS, a new text-to-speech system built for everything from live AI assistants to polished audio production. It arrives in two versions. Flash prioritizes speed and real-time interaction, while Plus leans toward higher-quality speech for professional use. That split matters. A customer service bot cannot afford an awkward pause before every response. An audiobook studio, meanwhile, may care far more about natural pacing, emotion and clarity than shaving a few milliseconds from generation time. Qwen Audio 3.0 TTS tries to serve both. Qwen Audio 3.0 TTS Goes Beyond…

Read More

NVIDIA is putting considerably more than a few graphics processors into South Korea’s AI ambitions. The chipmaker has partnered with the Korea Advanced Institute of Science and Technology, better known as KAIST, to establish a joint artificial intelligence research laboratory in Seoul. The new facility will focus on agentic AI models and systems designed around the Korean language, local industries and the country’s long-term technology needs. The collaboration carries an expected value of $300 million over its first five years. That includes annual computing contributions worth $50 million, giving KAIST researchers access to advanced NVIDIA AI infrastructure through local cloud…

Read More

Microsoft is no longer treating its homegrown AI models like side projects. The company has launched MAI-Image-2.5-Pro and MAI-Voice-2-Flash, two specialized models aimed at very different workloads. One focuses on high-fidelity image creation and editing. The other is built to generate natural speech quickly enough for busy call centers and real-time voice agents. Both models are now available in public preview through Microsoft Foundry. This is not simply another pair of models added to an already crowded catalog. Microsoft is steadily replacing outside technology with its own systems across Bing, PowerPoint, OneDrive, Dynamics 365 and Azure. That shift is becoming…

Read More

Black Forest Labs built its name around AI-generated images. With FLUX 3, the German AI company is attempting something much larger. Their newest project, the FLUX 3 multimodal AI model, aims to push the boundaries of artificial intelligence even further. The newly announced FLUX 3 multimodal AI model brings image generation, video, audio and action prediction into one architecture. Rather than stitching together separate systems for each type of media, Black Forest Labs says the model learns from them jointly. That difference matters. A video model that understands sound as part of the same scene could produce more natural timing…

Read More

Scandinavian Airlines is handing part of its airfare pricing process to artificial intelligence, and the early numbers have given the airline a reason to keep going. SAS is deploying Amadeus Air Pricing Optimization, known as AAPO, across its entire flight network. The system analyzes real-time information and adjusts ticket prices more continuously than traditional airline pricing models allow. In the first markets where SAS introduced the technology, the airline recorded an average revenue increase of more than 3%. Those initial results came from bookings made through the SAS website and channels connected through New Distribution Capability, or NDC. A few…

Read More

HubSpot is trying to clean up a problem that many businesses are only beginning to notice: once a company starts using several AI agents, those agents can quickly become just as fragmented as the software tools they were supposed to replace. The company has launched Agent Hub and Agent Builder in public beta for Professional and Enterprise customers. Together, the new tools give marketing, sales and customer service teams one central place to build, monitor and coordinate AI agents across HubSpot’s platform. It sounds like another AI dashboard at first. The more interesting part is the shared context underneath it.…

Read More

Abu Dhabi is bringing artificial intelligence into an area that rarely receives as much attention as chatbots, data centres or new language models: government policymaking. The Department of Government Enablement – Abu Dhabi has launched an AI-powered Policy Development Programme for executive directors and other senior government leaders involved in designing, implementing and overseeing public policy. This is not a general introduction to AI. The programme targets officials who make decisions, choose policy tools and deal with the problems that appear when government strategies move from presentation slides into the real world. AI Training Moves Closer to Government Decision-Making Much…

Read More

The UAE Ministry of Finance is getting measurable results from generative AI, not simply testing another chatbot behind the scenes. During the first half of 2026, the ministry’s call centre resolved 97.11% of customer enquiries during the first interaction. Its original target was 90%. Calls were also answered in an average of eight seconds, comfortably beating the ministry’s 20-second target. Those figures matter because government AI projects often arrive with ambitious announcements but very little operational evidence. Here, the ministry has attached numbers to the technology. Faster answers, fewer abandoned calls and less need for customers to contact the department…

Read More

Artificial intelligence may live inside software, but the industry’s biggest money is increasingly flowing into something far more physical. The recent MGX Aligned Data Centers acquisition is a key example of this trend toward investing in tangible infrastructure. Buildings. Power grids. Cooling systems. Land. Thousands of servers running around the clock. Abu Dhabi-backed investment firm MGX, the Artificial Intelligence Infrastructure Partnership and BlackRock’s Global Infrastructure Partners have completed their acquisition of Aligned Data Centers. The transaction gives Aligned an enterprise value of approximately $40 billion, placing it among the largest private digital infrastructure investments completed to date. The consortium is…

Read More

Google has officially put a name to its next major artificial intelligence model: Gemini 4. The confirmation came during Alphabet’s second-quarter 2026 earnings call, where CEO Sundar Pichai revealed that Google has started what he described as the company’s most ambitious pre-training run yet. That is the solid part. Gemini 4 exists, training has begun, and Google says it is already seeing encouraging progress at the frontier of AI development. Beyond that, the picture becomes less complete. Google has not announced a release date, benchmark results, pricing, model sizes or a detailed feature list. A lot of what people expect…

Read More

SpaceXAI may be preparing to take Grok Build far beyond its current role as an AI coding assistant. The new vision for Grok Build AI app platform could mean a broader set of features on the horizon. According to a new report, the company is developing features that could allow people to create applications with Grok Build, publish them directly, and share the finished product through a simple link. No separate hosting service. No awkward deployment process. Potentially, no need to leave the Grok ecosystem at all. Nothing has been formally announced yet, so the details remain tentative. Still, the…

Read More

Meta has spent the last few years putting AI almost everywhere it can. Chatbots. Image generation. Smart glasses. Now it is experimenting with something much quieter. The company is testing StoryKit, an AI-powered app designed to create personalized children’s stories. It is not another social media feature or AI assistant. Instead, StoryKit focuses on bedtime stories that parents can customize in just a few steps, complete with illustrations, narration, background music, and a moral woven into the story. The app is currently in limited testing, making it one of Meta’s latest AI experiments rather than a fully launched product. What…

Read More

Booking a flight is not usually the hardest part of a trip anymore. The harder part comes later. The unreadable menu. The train ticket machine that feels designed to punish tourists. The attraction with five ticket options and no clear explanation. The landmark you are staring at but do not fully understand. The delayed flight at midnight when nobody around speaks your language. Trip.com is trying to solve that awkward middle part of travel with a wider set of digital tools, many of them powered by TripGenie, its in-app AI travel assistant. This is not just about booking hotels faster.…

Read More

AI agents are starting to move out of the demo stage and into ordinary errands. That is the real story behind the new StepFun and Alipay partnership. At the 2026 World Artificial Intelligence Conference in Shanghai, foundation model company StepFun and Alipay announced a system-level partnership designed to connect AI agents with real-world daily services. Not just answering questions. Not just summarizing a page. The goal is to let users ask for something in plain language and have the agent coordinate the steps across services. That changes the pitch around AI agents a bit. For years, people heard that assistants…

Read More

AI has a data problem. Not just a copyright problem. Not just a bias problem. Something messier. A lot of what AI “knows” about people comes from images, captions, websites, and datasets that communities never approved, never reviewed, and sometimes never even saw. Then those same systems generate images that can flatten real lives into stereotypes. Microsoft is now testing a different approach. The company’s research team has introduced the Community Library Creator, a platform designed to let communities help define how they appear in AI-generated imagery. The idea is simple enough: instead of letting scraped internet data decide representation,…

Read More

ServiceNow just gave investors a fairly direct answer to one of the bigger questions hanging over enterprise software right now. In particular, ServiceNow AI growth is currently attracting attention as the company expands its artificial intelligence capabilities across the platform. Is AI going to weaken traditional software companies, or make the strongest ones even harder to replace? For ServiceNow, the latest quarter points toward the second answer. The company reported stronger-than-expected second-quarter 2026 results, with subscription revenue reaching $3.877 billion, up 24.5% from a year earlier. Total revenue came in at $3.987 billion, also up 24% year over year. Current…

Read More

Miramar is bringing artificial intelligence into youth creativity, and honestly, this is the kind of local AI story that deserves more attention. The city has launched an AI program for teens focused on creative projects like music, videos, and comics. Not corporate dashboards. Not another chatbot pitch. This one is about young people learning how to make things with AI. That matters because teenagers are already using AI tools. Some use them for homework. Some use them for ideas. Others are experimenting with images, songs, short videos, characters, and stories. Miramar seems to be leaning into that reality instead of…

Read More

Alibaba’s Qwen team has launched Qwen-Image-3.0, and this one is not just trying to make prettier AI pictures. That part matters. AI image tools have already become very good at producing polished visuals. Shiny product shots. Cinematic portraits. Fantasy landscapes. The usual stuff. Qwen-Image-3.0 seems aimed at a more annoying problem: making AI-generated images useful when the image contains lots of structure, text, diagrams, layouts, and tiny details that cannot fall apart halfway through. Alibaba describes Qwen-Image-3.0 as the third-generation foundational image generation model in the Qwen-Image series. The company says the new model is built around one idea: “Real.”…

Read More