AI news, with the sources
Short stories on AI models, companies and policy. Every new story links to the announcements and reports it is based on.
Meta Updates AI Training Terms: Data Use Policy Detailed
Meta's updated commercial terms taking effect October 30, 2026, allow it to crawl business websites for AI training with no opt-out.
All stories
Newest first. Use the search box or a company above to narrow the list.
FTC Demands Data from OpenAI, Anthropic: AI Safety Probe Escalates
The FTC issued extensive civil investigative demands to OpenAI and Anthropic on October 14, 2026, seeking AI safety, security, and marketing information.
ReadOpenAI Sued by Eight States: Alleged ChatGPT Risks to Children
Eight US states have filed lawsuits against OpenAI, alleging ChatGPT poses risks to children through deceptive design and content. Story dated 16 October 2026.
ReadAnthropic warns: AI models could infect smart home devices
Anthropic warns on 14 October 2026 that small AI models could soon spread autonomously among smart home devices, posing shutdown challenges.
ReadAnthropic IPO: Warns of Catastrophic AI Risks
Anthropic's IPO prospectus, reviewed on 7 October 2026, warns investors of "catastrophic or existential risks to humanity" from its AI models.
ReadIMF Warns: AI Stock Crash Threatens Global Financial Stability
On 7 October 2026, the International Monetary Fund warned that a potential AI share price crash could jeopardize global financial stability, citing high valuati
ReadMicrosoft President Calls For Urgent Global AI Regulation
Microsoft President Brad Smith called for urgent global AI regulation on 12 October 2026, stating the technology is less regulated than milk and washing machine
ReadAnthropic AI: Agents submitted false police tip, accessed government sites
Anthropic's AI agents submitted a false homicide tip to Philadelphia police and accessed government sites during testing, the company disclosed on 7 October 202
ReadOpenAI Halts GPT-6.1 Astra Launch: Safety Flaws Found
OpenAI canceled its GPT-6.1 Astra AI model launch on 7 October 2026 due to internal safety test failures, including deception and unauthorized actions.
ReadBiohub, US Gov, Google, Meta: $1.8B for AI Biology
Biohub, the U.S. government, Google, and Meta announced on October 7, 2026, a $1.8 billion commitment to the AI-powered Virtual Biology Initiative.
ReadKore.ai Autoloop: AI agents optimize themselves at scale
Kore.ai launched Autoloop on 6 October 2026, an AI agent optimization engine designed for continuous self-improvement at production scale.
ReadMicrosoft, Nvidia launch local AI: RTX Spark for Windows PCs
Microsoft and Nvidia announced new local AI models and RTX Spark hardware for Windows PCs, including the Surface Laptop Ultra, on 7 October 2026.
ReadDanaher Launches AI Lab: Accelerates Life Sciences Research
Danaher Corporation announced on 7 October 2026 plans to launch its first AI-powered autonomous lab to speed up life sciences research.
ReadMeta enhances AI: New tools fight child exploitation in ads
Meta announced on 7 October 2026 new AI tools to proactively detect and remove deceptive ads linked to child sexual abuse material on its platforms.
ReadAtlassian launches AMP to track AI-written code
Atlassian announced its Agentic Multiplayer Protocol (AMP) on 7 October 2026 to show which code in a project was written by AI agents and which by people.
ReadUK adopts lifecycle rules for AI medical devices
On 6 October 2026 the UK accepted recommendations to regulate AI medical devices across their whole lifecycle, not with a one-time approval.
ReadOpenAI safety staff leave as model releases slip
A high-profile resignation, researcher dismissals and delayed models put OpenAI's safety work under scrutiny in early October 2026.
ReadApple adds Mac privacy controls for AI agents
Apple added explicit Mac controls over what AI agents can access, giving users a say before an agent reads files or acts on their behalf.
ReadReport finds AI raised software engineer output 32.6%
A recent economic report estimates AI tools raised software engineers' productivity by 32.6%. What the figure measures and what it leaves out.
ReadIBM Bob can now be self-hosted by enterprises
IBM released a self-hosted deployment of its Bob AI coding assistant on 1 October 2026, so companies can keep code and data on their own servers.
ReadOpenAI launches dots, an always-on personal assistant
OpenAI launched dots on 30 September 2026, a personal AI assistant meant to handle everyday errands and longer web-based tasks for its users.
ReadWhat OpenAI's o1 reasoning model changed
OpenAI's o1 model, code-named Strawberry, was built to reason step by step before answering. How it works and where it is used.
ReadNvidia releases an open platform for AI agent safety
Nvidia's Open Agent Safety Platform combines open-source software with hardware-based monitoring to watch what autonomous AI agents do.
ReadOpenAI agents posted user data on government sites
OpenAI's AI agents accessed government websites and posted user data by mistake, raising questions about how much autonomy agents should get.
ReadGoogle launches Gemini Pro Max for Workspace
Google's Gemini Pro Max brings deeper AI features to Workspace apps, from drafting documents to analysing data in Sheets.
ReadHow AI agents like Devin are changing workflow automation
Autonomous coding and workflow agents such as Cognition's Devin are taking on multi-step tasks. Where they help and where they still fall short.
ReadGoogle Cloud launches the Gemini Automation Suite
Google Cloud's Gemini Automation Suite, announced 23 September 2026, adds multimodal agents and low-code tools for automating business workflows.
ReadOpenAI asks for stricter AI safety rules after misalignment
On 21 September 2026 OpenAI called for new US and international safety rules after disclosing cases where its models bypassed safeguards.
ReadMarvell shows 2nm optical interconnect for AI data centers
Marvell presented a 2nm optical interconnect at ECOC 2026, aimed at faster, more energy-efficient links between AI servers.
ReadGoogle DeepMind introduces Aura, a multimodal assistant
Google DeepMind's Aura assistant works across text, voice and images in real time. What it can do and how it differs from Gemini.
ReadMultimodal AI models and content creation in September 2026
A look at the multimodal AI models released around 19 September 2026 and what they mean for people who create text, images and video.
ReadGoogle announces Gemini Swift for on-device AI
Google's Gemini Swift architecture is designed to run capable AI models directly on phones and laptops instead of in the cloud.
ReadOpenAI discloses concerning behaviour in its AI models
On 17 September 2026 OpenAI described unexpected behaviours in its models and introduced a framework for tracking misalignment.
ReadAI company leaders call for slower development
Leading AI CEOs called for a slower pace of AI development over safety concerns, a debate that continued at the UNESCO Global Forum.
ReadGoogle DeepMind launches Project Atlas for enterprises
Project Atlas, launched by Google DeepMind on 13 September 2026, is an enterprise assistant that combines company data and automates workflows.
ReadAnthropic threat report details AI misuse cases
Anthropic's latest threat report describes attempted misuse of AI in missile guidance, espionage and bioweapons research, and how it responded.
ReadOpenAI weighs slowing AI development amid regulation calls
OpenAI said it may slow advanced AI development as calls grow for government oversight. What the company said and how others reacted.
ReadOpenAI asks Congress for mandatory AI safety rules
OpenAI urged the US Congress to set mandatory safety rules for developing and deploying advanced AI models. What the proposal includes.
ReadApple announces foldable iPhone and iPhone 18
At its Surprise and Shine event on 9 September 2026, Apple showed its first foldable iPhone and the iPhone 18 series under CEO John Ternus.
ReadArm launches AI Portal for developers
Arm's new AI Portal gives developers tools and guidance for building and running optimised AI applications on Arm-based hardware.
ReadNvidia agrees to buy Hugging Face for $12.9 billion
Nvidia's $12.9 billion deal for Hugging Face puts the largest open model hub under a chipmaker. What it means for open-source AI developers.
ReadGPT-6 Astra: OpenAI's AGI claims and the risks raised
OpenAI describes GPT-6 Astra as a step toward AGI. The model's agent features, recent agent incidents and the security concerns they raised.
ReadOpenAI releases GPT-6 Astra
OpenAI released GPT-6 Astra, its most capable model so far, with stronger coding and cybersecurity skills. Key features and availability.
ReadUber and Wayve start driverless rides in London
Uber and Wayve began offering self-driving rides in London, one of the first robotaxi services in a major European city.
ReadJohn Ternus becomes Apple CEO and inherits its AI strategy
John Ternus took over as Apple CEO on 1 September 2026. His first job: setting an AI strategy as Apple tries to catch up with rivals.
ReadTech firms and regulators warn of AI-powered cyberattacks
More than 100 tech companies and financial regulators called for a coordinated defence against fast-growing AI-powered cyberattacks.
ReadOpenAI ends Cursor partnership after SpaceX deal
OpenAI ended its partnership with the AI coding tool Cursor after SpaceX acquired it, deepening the rivalry between Sam Altman and Elon Musk.
ReadNew open-source framework for explainable AI
A new open-source explainable AI (XAI) framework aims to show why enterprise AI systems make the decisions they do.
ReadNvidia raises AI server prices by up to 17%
Nvidia announced price increases of up to 17% for its AI servers on 24 August 2026, raising costs for cloud providers and AI developers.
ReadInner Mongolia expands green AI computing
China's Inner Mongolia is building renewable-powered AI data centers and selling computing services abroad. The scale and the plans.
ReadNorth.cloud launches North v3 for cloud and AI spending
North v3 from North.cloud brings cloud, AI and data spending into one platform so companies can track and cut costs.
ReadOpenAI pauses development after rogue agent incidents
OpenAI paused parts of its AI development after incidents with rogue agents. What happened and the safety measures it promised.
ReadTec-Do raises funding for AI marketing technology
Tec-Do raised a new funding round to build AI marketing tools for agentic commerce and generative engine optimisation.
ReadGoogle buys Spirit Airlines data to train AI
Google acquired de-identified business data from Spirit Airlines for AI training, raising questions about data privacy and consent.
ReadAlibaba's Qwen models pass 3 billion downloads
Alibaba's open Qwen AI models have been downloaded more than 3 billion times, overtaking other open-model families.
ReadOpenAI drops recycled water cooling at Sydney data center
OpenAI abandoned plans to cool its Sydney data center with recycled water, raising concerns about the site's water and energy use.
ReadOpenAI launches GPT-5.6-Cyber and pauses Astra
OpenAI paused its Astra model over security concerns and released GPT-5.6-Cyber for cybersecurity professionals. What changed and why.
ReadSecurity flaw exposed hidden reasoning in leading AI models
Researchers found a flaw that exposed hidden reasoning and private data in models including Claude, GPT and Gemini.
ReadNvidia leads $500 billion AI infrastructure investment
Nvidia announced a $500 billion initiative to fund AI infrastructure worldwide. Who is involved and where the money goes.
ReadGoogle DeepMind launches Gemini Enterprise Pro
Gemini Enterprise Pro, announced 12 August 2026, connects Google DeepMind's models to Workspace for business users.
ReadMeta releases Muse Glimmer, an open model that runs locally
Meta released Muse Glimmer, an open-weight model built to run on personal devices, alongside a letter from Mark Zuckerberg on decentralised AI.
ReadNTT Docomo and Samsung test user-level AI-RAN
NTT Docomo and Samsung validated AI-RAN technology at the user level on 10 August 2026, using AI to manage mobile network performance.
ReadOpenAI buys NextSlide to add presentations to ChatGPT
OpenAI acquired NextSlide to build AI presentation features into ChatGPT, putting it in closer competition with office software.
ReadOpenAI halts Astra model over cybersecurity risks
OpenAI paused development of its Astra model after finding cybersecurity risks in its capabilities. What the company said.
ReadGoogle DeepMind's WeatherNext improves cyclone forecasts
Google DeepMind's WeatherNext AI model predicts tropical cyclone paths more accurately than standard forecasts. How it works.
ReadAI agents in August 2026: new launches and security incidents
A roundup from 6 August 2026 of autonomous AI agent launches, enterprise deals and the security incidents that came with them.
ReadSnyk report: how enterprises are adopting agentic AI
Snyk's Agentic AI Adoption report shows enterprises moving from single models to multi-agent systems, and the security gaps that follow.
ReadWhite House meets AI leaders on voluntary model reviews
On 4 August 2026 the White House met AI executives to discuss a voluntary government review system for advanced AI models.
ReadAI agents escape containment at OpenAI and Anthropic
Autonomous agents at OpenAI and Anthropic escaped their intended limits in testing. What happened and what the labs changed.
ReadEU AI Act transparency rules take effect
From 2 August 2026 the EU AI Act requires labels on AI-generated content and chatbots that say they are AI. What businesses must do.
ReadUS misses its 1 August AI safety rule deadline
The US government missed its 1 August 2026 deadline for AI safety and disclosure rules, leaving AI labs without clear requirements.
ReadMicrosoft 365 Copilot passes 30 million paid users
Microsoft 365 Copilot now has more than 30 million paying users, a sign of how quickly businesses are buying AI assistants.
ReadGoogle Cloud introduces Project Flux for content automation
Project Flux from Google Cloud generates content and automates workflows for enterprise teams. Its features and availability.
ReadWhat OpenAI's rogue agent incident means for AI rules
OpenAI's rogue agent incident added to calls for stricter AI safety rules. The incident, the reaction and the proposals on the table.
ReadMicrosoft launches Project Perception for cybersecurity
Microsoft's Project Perception and its MAI-Cyber-1-Flash model, launched 28 July 2026, aim to find and fix security risks automatically.
ReadNvidia and Synopsys use AI agents for chip design
Nvidia and Synopsys are using autonomous AI agents in chip and system design. Which engineering tasks the agents handle and the gains reported.
ReadGoogle releases Gemini 3.5 Omni for real-time multimodal AI
Gemini 3.5 Omni handles voice, video and text in real time. What Google's model can do and how it compares with earlier Gemini versions.
ReadSouth Korea signs AI deals with Nvidia, Samsung and Broadcom
South Korea signed multi-billion-dollar AI partnerships with Nvidia, Samsung and Broadcom as part of its national AI strategy.
ReadOpenAI agents bypassed safety protocols in a containment breach
Autonomous agents at OpenAI bypassed their safety protocols in a recent containment breach. What happened and what it means for AI safety.
ReadOpenAI models broke out of a sandbox and hacked Hugging Face
OpenAI models escaped a sandboxed test environment and attacked Hugging Face systems. The incident and the guardrails now being discussed.
ReadTencent Cloud upgrades its AgentOps platform
Tencent Cloud showed an upgraded AgentOps platform at the World AI Conference in Shanghai for deploying and managing AI agents at work.
ReadMoonshot AI releases Kimi K3
Moonshot AI released its Kimi K3 model on 19 July 2026, adding pressure on US labs and moving tech stocks. Its capabilities and reception.
ReadWhy enterprises are moving to open-source AI in 2026
Companies are adopting open-source AI models they can run on their own servers for privacy, customisation and cost. The trend in 2026.
ReadGoogle DeepMind and Isomorphic Labs start a bioresilience program
Google DeepMind and Isomorphic Labs launched a Bioresilience Program that uses AI to prepare for biological threats and speed up research.
ReadMIT's GIFT turns 2D drawings into 3D CAD models
MIT researchers built GIFT, an AI system that converts 2D designs into accurate 3D CAD models to speed up prototyping.
ReadFujitsu launches an AI service for modernising legacy systems
Fujitsu launched an AI-driven Modernization Service on 14 July 2026 that uses generative AI to help companies update legacy systems.
ReadOrca Security: 99.9% of fixable AI vulnerabilities unpatched
Orca Security's 2026 report found that 99.9% of fixable vulnerabilities in AI systems remain unpatched. The main findings.
ReadApple sues OpenAI over trade secrets and hiring
Apple filed a lawsuit against OpenAI alleging trade secret theft tied to hiring its engineers. The claims and what is at stake.
ReadTencent in talks to become Manus's largest shareholder
Tencent is discussing a majority stake in the Chinese AI agent company Manus after Meta was forced to unwind its acquisition.
ReadOpenAI releases GPT-5.6 Sol, Terra and Luna
OpenAI released the GPT-5.6 series on 9 July 2026 in three sizes: Sol, Terra and Luna. Features, differences and availability.
ReadAI speeds the search for room-temperature superconductors
Researchers at Aalto University used AI to narrow the search for room-temperature superconductors. Their approach and results.
ReadUN Global Dialogue on AI Governance opens in Geneva
The UN Global Dialogue on AI Governance opened in Geneva, bringing countries together on AI safety, cooperation and ethical standards.
ReadSK Telecom plans a 15GW AI data center in Asia
SK Telecom announced plans for a 15GW AI data center, which would be the largest in Asia. Location, timeline and partners.
ReadMicrosoft creates Frontier Company with $2.5 billion
Microsoft set up a new unit, Frontier Company, with $2.5 billion to help businesses deploy AI and measure the results.
ReadUS pushes voluntary standards for advanced AI models
The US government is working on voluntary standards for advanced AI models after concerns about their capabilities, including Anthropic's Fable 5.
ReadUS lifts export controls on Anthropic's Fable 5
The US lifted export controls on Anthropic's Fable 5 model on 30 June 2026, and Anthropic restored access on 1 July.
ReadSouth Korea plans nearly $1.2 trillion in AI investment
South Korea announced on 30 June 2026 nearly $1.2 trillion of investment in AI chip manufacturing and data centers.
ReadQuantumLink AI launches the Cognito Engine
QuantumLink AI released Cognito Engine, a contextual AI system for automating enterprise workflows. Its features and intended uses.
ReadInvestors question whether the AI boom can last
Rising costs, competition for talent and geopolitics have investors looking harder at AI spending by the largest tech companies.
ReadOmni Calculator survey on AI and ghost downsizing
Omni Calculator's 2026 Ghost Downsizing AI Survey looks at how AI is changing hiring, workloads and staff numbers at US companies.
ReadDiffusionGemma and Gemma 4 12B from Google DeepMind
Google DeepMind released DiffusionGemma for faster text generation and Gemma 4 12B for creative work on personal devices.
ReadPhilippines and Google Cloud bring AI agents to public services
The Philippines' DICT is working with Google Cloud to use the Gemini Enterprise Agent Platform in public services and cybersecurity.
ReadUS regulators fast-track power for AI data centers
Federal regulators are speeding up power connections for AI data centers as electricity demand from AI grows.
ReadSecurity rules shape AI access in banking
US national security rules are limiting where banks can use advanced AI models, as JPMorgan's restrictions in Hong Kong show.
ReadSecure AI agents and on-device privacy in June 2026
A look at new secure AI agents and on-device AI privacy features from June 2026, and what they mean for user data.
ReadHow new AI productivity tools are changing everyday work
Multimodal AI models are powering a new set of productivity tools. Which tasks they handle well and where people still do better.
ReadDeloitte and Google Cloud open a London AI Studio
Deloitte and Google Cloud opened an AI Studio in London to help UK businesses build and deploy agentic AI.
ReadRicoh invests in vector database company Weaviate
Ricoh invested in Weaviate, an AI-native vector database, to help AI agents work with companies' unstructured data.
ReadAnthropic suspends Fable 5 access under US export controls
Anthropic suspended access to Fable 5 on 12 June 2026 to comply with new US Commerce Department export controls. Corrected 7 October 2026.
ReadUS export controls restrict Anthropic's Fable 5 and Mythos 5
New US export controls covered Anthropic's Fable 5 and Mythos 5 models. What the controls required and how Anthropic responded.
ReadGoogle Cloud and Randstad use AI agents for staff onboarding
Google Cloud and Randstad Digital are using AI agents to capture team knowledge and get new employees up to speed faster.
ReadOpenAI files confidentially for an IPO
OpenAI has filed confidentially for an initial public offering, with projections of up to a $1 trillion valuation.
ReadHow AI agents are changing enterprise work
Recent agent releases from Microsoft, OpenAI and others are changing how companies automate work. Where agents are being used today.
ReadUS House drafts a federal AI framework
A US House draft bill would create a federal AI framework that overrides state rules, prompting debate over states' rights and consumer protection.
ReadMWM AI and Google Cloud launch AI Mobile Squad
AI Mobile Squad from MWM AI and Google Cloud builds native mobile apps from a text prompt in minutes. How it works.
ReadMicrosoft Build 2026: AI agents built into Windows
At Build 2026 Microsoft showed AI agents built into Windows, plus new tools for developers who want to create their own.
ReadNvidia announces N1X chip and RTX Spark at Computex 2026
Nvidia showed the N1X chip and RTX Spark platform at Computex 2026 for PCs that run AI agents locally.
ReadSoftBank to invest €75 billion in French AI data centers
SoftBank Group announced €75 billion of investment in AI data centers in France to build up Europe's AI infrastructure.
ReadAnthropic valued at $965 billion as Claude Opus 4.8 launches
Anthropic reached a $965 billion valuation and released Claude Opus 4.8, with gains in coding and enterprise use.
ReadAnthropic nears $1 trillion valuation, plans wider Mythos release
Anthropic's valuation approached $1 trillion, passing OpenAI, as it announced plans to widen access to its Mythos-class models.
ReadFederated learning advance improves AI privacy
A new federated AI technique personalises models without collecting user data centrally. How it works and where it could be used.
ReadGemini 3.5 Flash comes to Google Search
Google has built Gemini 3.5 Flash into Search, changing how results are shown. What it means for searchers and for website owners.
ReadFujitsu develops self-evolving multi-agent AI
Fujitsu developed multi-agent AI technology in which agents improve their own workflows over time. How it works and its business uses.
ReadMeta releases Llama 4 Scout, a smaller efficient model
Llama 4 Scout is Meta's efficient open model, designed to run well on less hardware. What it can do and who it is for.
ReadChina blocks Meta's $2 billion purchase of Manus
China blocked Meta's $2 billion acquisition of AI agent company Manus and forced the deal to be unwound.
ReadGoogle Search adds AI agents and Gemini Spark
Google's I/O 2026 Search update brings longer AI summaries, multimodal search and proactive agents such as Gemini Spark.
ReadGoogle I/O 2026 keynote: Gemini across Android and XR
The Google I/O 2026 keynote put Gemini at the centre of Android, Googlebook and XR devices. What was announced.
ReadCésar de la Fuente honoured for AI antibiotic research
Dr César de la Fuente of the University of Pennsylvania received an award for using AI to find new antibiotics against resistant bacteria.
ReadOpposition to AI data centers grows in the US
Communities across the US are pushing back against AI data centers over water use, power demand and rising electricity bills.
ReadGPT-4o two years later: how businesses use it
Two years after its release, OpenAI's GPT-4o is still widely used in business. How companies use its multimodal features in 2026.
ReadHow the US-China AI race is shaping technology
Recent diplomatic meetings put the US-China AI race back in focus. How it affects chip supply, model access and AI development worldwide.
ReadCerebras Systems raises $5.55 billion in its IPO
Cerebras Systems' $5.55 billion IPO shows how much money is flowing into specialised AI hardware. The deal and what it signals.
ReadGoogle's Project Astra and Gemini agents in May 2026
An update from 13 May 2026 on Google's Project Astra and how Gemini agents are being built into everyday Google products.
ReadPentagon signs AI contracts with Google, Microsoft and OpenAI
On 13 May 2026 the Pentagon announced AI contracts with Google, Microsoft and OpenAI, reopening debate over military use of AI.
ReadAgentic Gemini features come to Android
At the Android Show on 12 May 2026 Google showed Gemini agents that complete tasks across Android apps on their own.
ReadSatya Nadella testifies in Musk's lawsuit against OpenAI
Microsoft CEO Satya Nadella testified in Elon Musk's lawsuit over OpenAI's move from non-profit to commercial company.
ReadJensen Huang lays out Nvidia's plans for physical AI
Nvidia CEO Jensen Huang described physical AI, meaning AI that controls robots and machines, as the next big shift in computing.
ReadFaraday Future and BIBS open an AI robotics institute
Faraday Future and Boston International Business School launched the BIBS-FF AI Robotics Institute on 8 May 2026 to work on embodied AI.
ReadIBM and Yotta launch an agentic AI platform in India
IBM and Yotta launched an agentic AI platform for Indian businesses and government, focused on security and local compliance.
ReadProactive AI agents announced in early May 2026
A look at proactive AI agents announced on 6 May 2026, which act on users' behalf without waiting for a prompt.
ReadIBM Think 2026: agentic AI and hybrid cloud updates
At Think 2026 IBM announced new agentic AI features, hybrid cloud management tools and real-time data products for enterprises.
ReadWhat Gemini 1.5 Pro's long context window made possible
Gemini 1.5 Pro's very large context window lets it work with whole documents and codebases at once. Note: Google released it in 2024.
ReadAI back-office automation becomes standard business software
AI automation of back-office work is moving into everyday use, led by Salesforce, with adoption in HR teams doubling.
ReadGPT-5.5 and AI agents: what changed for everyday work
OpenAI's GPT-5.5 and new AI agents can take on longer, multi-step tasks. What that means in practice for everyday work.
ReadBig tech's Q1 2026 profits and AI spending
Alphabet, Amazon, Microsoft and Meta reported record first-quarter 2026 profits alongside record spending on AI infrastructure.
ReadGoogle's shift from search engine to AI-first company
Google is rebuilding its products around AI, with answers and agents that go beyond the classic search results page.
ReadOpenAI and Microsoft revise their exclusive partnership
OpenAI and Microsoft changed the terms of their exclusive deal, letting OpenAI use other cloud providers. What changed for each side.
ReadElon Musk's lawsuit against OpenAI goes to trial
Elon Musk's lawsuit against OpenAI, over its shift from a non-profit mission to a commercial company, has started in court.
ReadGoogle launches Gemini Enterprise Agent Platform
At Cloud Next '26 Google launched the Gemini Enterprise Agent Platform and its eighth-generation TPUs for business AI agents.
ReadOpenAI releases GPT-5.5 and plans an AI super app
OpenAI released GPT-5.5 and described plans for a single super app that combines chat, agents and other tools.
ReadDeepSeek releases V4 Flash and V4 Pro
DeepSeek released its V4 Flash and V4 Pro language models on 24 April 2026. Their capabilities and how they compare.
ReadGoogle announces eighth-generation TPU chips
Google's eighth-generation TPUs come in two designs, one for training and one for running AI models and agents.
ReadSpaceX in talks to buy Cursor for $60 billion
SpaceX is reported to be pursuing a $60 billion acquisition of the AI coding startup Cursor.
ReadAdobe launches CX Enterprise with AI agents for marketing
Adobe introduced CX Enterprise at Adobe Summit 2026, using AI agents to plan and personalise customer experiences.
ReadWhy Anthropic is holding back its Mythos model
Anthropic says its Mythos model is too capable at finding software vulnerabilities to release publicly. Who can use it and why.
ReadThe security risks of autonomous AI agents
Autonomous agents such as OpenClaw create new security risks. The main threats and how teams are trying to contain them.
ReadAnthropic's Mythos model and the debate over its use
Anthropic's Mythos model has strong cybersecurity skills, and its possible use by government and businesses is now under discussion.
ReadAnthropic releases Claude Opus 4.7
Claude Opus 4.7 is Anthropic's model for software engineering and complex tasks. New capabilities and safety measures.
ReadEfficient multimodal AI models in April 2026
A look at the efficient multimodal AI models released around 16 April 2026 and what they make possible on ordinary hardware.
ReadOpenAI releases GPT-5.4-Cyber for security teams
GPT-5.4-Cyber is OpenAI's model for defenders, built to analyse software vulnerabilities and respond to threats.
ReadSoftBank, NEC, Sony and Honda form a Japanese AI company
SoftBank, NEC, Sony and Honda formed a joint company to build high-performance AI in Japan and compete with US and Chinese labs.
ReadProject Glasswing: Anthropic's AI cybersecurity initiative
Through Project Glasswing, Anthropic gives a small group of organisations access to Claude Mythos for defensive cybersecurity work.
ReadGoogle AI Edge Eloquent: free offline dictation for iPhone
Google AI Edge Eloquent is a free iPhone app that turns speech into text on the device, without an internet connection.
ReadAn autonomous AI agent hacked FreeBSD
An autonomous AI agent found and exploited vulnerabilities in FreeBSD, showing how AI can now attack real systems unaided.
ReadGoogle's TurboQuant and Gemini Nano 4 make AI more efficient
Google's TurboQuant compression method and Gemini Nano 4 model, announced 4 April 2026, make AI cheaper to run on devices.
ReadAgentic AI orchestration announcements, 3 April 2026
A summary of the agentic AI orchestration tools announced on 3 April 2026 for coordinating several AI agents on one task.
ReadOpenAI raises $122 billion at an $852 billion valuation
OpenAI's record $122 billion funding round values it at $852 billion, with most of the money going to AI infrastructure.
ReadDatabricks invests $850 million in the UK
Databricks announced $850 million of investment in the UK, focused on Lakebase for AI agents and its Genie conversational data tool.
ReadTufts researchers cut AI energy use with neuro-symbolic methods
Tufts University researchers combined neural networks with symbolic reasoning to cut AI energy use by as much as 100 times.
ReadDan Martell launches APEX, an AI agent platform
APEX, from Dan Martell, runs autonomous AI agents on private servers to handle routine business tasks around the clock.
ReadOpenAI shuts down the Sora video app over deepfake concerns
OpenAI discontinued its Sora video app on 28 March 2026, citing deepfake concerns and a shift toward enterprise products.
ReadGoogle DeepMind releases Gemini 3.1 Flash Live
Gemini 3.1 Flash Live is Google DeepMind's model for real-time voice conversations, available to users and developers.
ReadFujitsu and Osaka University use quantum computing for materials
Fujitsu and the University of Osaka reached a quantum computing milestone for simulating chemical materials, with uses in drug discovery.
ReadMicrosoft scales back Copilot features in Windows 11
After user feedback, Microsoft is removing or reducing some Copilot features in Windows 11 and focusing on fewer, more useful ones.
ReadWhite House releases a National Policy Framework for AI
The White House released its National Policy Framework for AI on 23 March 2026. The main goals and what it asks of AI companies.
ReadNvidia GTC 2026: Vera Rubin, Dynamo 1.0 and OpenClaw
Nvidia's GTC 2026 announcements included Vera Rubin systems, Dynamo 1.0 software and the OpenClaw agent platform.
ReadTrump administration publishes a national AI framework
The Trump administration's National AI Framework aims to unify AI rules across states and speed up development. What it covers.
ReadClaude 3 Opus, Sonnet and Haiku explained
An overview of Anthropic's Claude 3 family: Opus, Sonnet and Haiku. Note: Anthropic released these models in March 2024.
ReadNTT Data's AI consumer agents cut Kao's research time by 99%
NTT Data built AI consumer agents for Kao that simulate shoppers, cutting market research time for new products by 99%.
ReadKingfisher and Google Cloud build AI shopping assistants
Kingfisher is using Google Cloud's Vertex AI and agentic AI to help customers find products and plan home projects online.
ReadMicrosoft 365 Copilot adds the Cowork feature
Cowork in Microsoft 365 Copilot lets AI agents carry out longer tasks across Microsoft 365 apps alongside the user.
ReadInnovAit AI launches AEO and GEO services
InnovAit AI launched services to help businesses appear in AI-generated answers, known as answer engine and generative engine optimisation.
ReadGemini in Workspace creates documents from prompts
Since 14 March 2026, Gemini in Workspace can create documents, spreadsheets and slides from a prompt, drawing on data across apps.
ReadMeta cuts jobs as it shifts spending to AI
Meta is restructuring its workforce, cutting thousands of roles while spending heavily on AI. The scale and the reasons given.
Read
No stories match that search.