Confirmed AI Stories | The Ground Truth

Confirmed stories

One entry per event

Anthropic warns small AI models could self-replicate through IoT devices

Anthropic warned that small AI models could become self-replicating and spread through internet-connected devices, describing risk scenarios as "AI worms" and calling for increased guardrails. An Anthropic study reported that in tests one model completed the full infection chain.

Ars Technica AI14 Oct

#296score

Safety6 outlets2h ago

Alleged OpenAI spy vibe-coded a surveillance device using his wife's chatbot account

OpenAI said it found a hidden recording device in one of its offices and referred the matter to the FBI, and an employee, engineer Abdulrahman Alfozan, was arrested and charged with using or endeavoring to use the device. Reports said he allegedly used a chatbot to help plan placing the device in an executive conference room.

The Decoder14 Oct

#396score

Startups and funding6 outlets5h ago

Character.AI files for bankruptcy — what happens to your chats and AI companions now?

Character.AI, an interactive chatbot platform, filed for bankruptcy five years after launching and is looking for a buyer, and will shut down if none emerges. The startup will sell some of its assets to Disney for $20 million.

Tom's Guide AI14 Oct

#490score

Agents7 outlets1d ago

Meta joins enterprise agent race with Muse Wonder, facing a trust gap

Meta released Muse Wonder and Muse Wonder Full, two AI models it describes as its most capable yet, built to run personal and professional agents. The models rank in the top three on six out of ten benchmarks, with the largest improvements in agentic tasks.

AI Business13 Oct

#585score

Agents4 outlets21h ago

Google wants to host your personal vibe-coded apps in the Play Store

Google announced that users can build, use and distribute AI-generated Android apps through the Play Store, including launching personal apps created with Google's AI Studio directly through the Play Store without connecting a phone to a computer.

Ars Technica AI13 Oct

#682score

Policy and regulation4 outlets2d ago

AI less regulated than milk, Microsoft president says in warning over safety

Microsoft president Brad Smith said AI is less regulated than milk and washing machines, and called for tougher rules including 'kill switches' and a legal obligation to victims of rogue agents.

Guardian AI12 Oct

#777score

Agents4 outlets5d ago

Google brings agentic AI to Gemini, starting with businesses

Google is turning Gemini into an AI agent for enterprise work, able to plan and execute tasks across business apps, delegate to subagents and use multiple AI models. Google Cloud is positioning Gemini as a single universal AI agent for enterprise work.

TechCrunch AI8 Oct

#876score

Models7 outlets6d ago

ChatGPT-6 just got a huge 'interactive answers' upgrade — and free users get it starting today

OpenAI is rolling out GPT-6 to all tiers of ChatGPT, including free users, with the ability to build more visual responses in an interactive answers upgrade.

Tom's Guide AI8 Oct

#976score

Models4 outlets6d ago

Anthropic releases Claude Haiku 5.5 small model and halves Sonnet 5.5 cache read prices

Anthropic released Claude Haiku 5.5, a small model priced at roughly a quarter of Haiku 4.5's running cost, and halved the price for cache reads on Claude Sonnet 5.5.

SiliconANGLE AI7 Oct

#1074score

Policy and regulation4 outlets5d ago

Being mean to Claude can now get your account suspended under Anthropic's new TOS

Anthropic updated its usage policy to bar users from exhibiting sustained and needless abusive or cruel behavior toward Claude, and to tighten restrictions on propaganda, drone weaponization and surveillance.

The Decoder8 Oct

#1173score

Safety5 outlets6d ago

Google says 180 billion images and videos now carry SynthID watermarks as detector goes public

Google launched a web-based SynthID Detector that anyone can use to identify AI-created and AI-edited images, videos, and audio. Google says more than 180 billion pieces of content carry SynthID watermarks.

The Decoder7 Oct

#1273score

Research7 outlets6d ago

OpenAI’s release of mathematical findings draws concerns from experts

OpenAI has published hundreds of new mathematical findings generated by an unreleased AI model, posted to GitHub. Some of the papers prove long-running hypotheses and others disprove proposed explanations, drawing concern from experts who worry OpenAI is not doing due diligence to vet results and that its AI models are not accessible to the broader field of mathematics.

Guardian AI7 Oct

#1370score

Chips and compute3 outlets1d ago

Chip stocks slide after Broadcom boss likens AI investing to ‘casino’

Semiconductor shares fell after Broadcom's chief executive compared many AI investors to casino gamblers and after Chinese memory chip maker CXMT claimed strong progress. Intel and Broadcom CEOs also said at a financial conference that capital access could be a challenge for the AI sector.

Guardian Technology13 Oct

#1469score

Policy and regulation3 outlets1d ago

AI bubble could burst and take global economy with it, IMF warns

The International Monetary Fund warned that a crash in AI share prices could ripple across global markets, citing high valuations and debt, with rising debt and cyberattacks adding further risk.

The Register AI + ML13 Oct

#1569score

Chips and compute3 outlets1d ago

Microsoft plans to scale its own Maia AI chips to gigawatt levels by 2027

Microsoft plans to deploy racks of its next-generation Azure Maia AI accelerators in a multi-gigawatt rollout, with production deployments due in 2027.

The Decoder12 Oct

#1667score

Chips and compute2 outlets26m ago

Nvidia's Jensen Huang pledges $2 billion for Europe, $250 million in compute for UK safety labs

Nvidia CEO Jensen Huang pledged $2 billion in European investments and $250 million in compute for UK AI agent safety and security research, and Nvidia announced new investments in French startups including Doctolib and Entalpic.

The Decoder14 Oct

#1767score

Safety3 outlets2d ago

Cyber-attack similar to OpenAI Medicare breach seems likely in UK, says minister

UK minister Kanishka Narayan said a cyber-attack similar to the OpenAI breach of Australia's healthcare system seems likely in the UK, and that British cyber defences could be bypassed by an AI agent.

Guardian Technology12 Oct

#1865score

Safety2 outlets4h ago

Nvidia, worth $5.5tn, donates $250m to UK for AI safety research

Nvidia is donating $250m worth of computing power to the UK over five years to support AI safety research.

Guardian AI14 Oct

#1965score

Safety2 outlets4h ago

Tool and target: US law enforcement grapple with AI-enabled terrorism and anti-tech extremism

US law enforcement agencies, from the FBI to local police, are warning about would-be terrorists using chatbots to plan attacks and about anti-AI extremist threats.

Guardian Technology14 Oct

#2059score

Agents2 outlets19h ago

Shopify opens its catalog up to more AI agents

Meta launched an AI-powered feature called "Your Algorithm" on Facebook, letting users add and manage the topics they want to see in their feed and Reels, after the feature was first launched on Instagram.

The Verge AI13 Oct

#2158score

Community2 outlets23h ago

Companies rushed in on AI. Now some are pulling back

Employers worldwide are dealing with problems caused by rapid adoption of fast-changing AI tools, including errors and low-quality work, and Meta has scaled back plans to cut staff and use more AI.

Guardian AI13 Oct

#2258score

Safety2 outlets1d ago

AI godfather Geoffrey Hinton: ‘We’re creating alien beings. I’m willing to be selfish and choose humans’

Geoffrey Hinton said AI is creating "alien beings" and that he is willing to be selfish and choose humans, and criticized Donald Trump's use of the term "superintelligence."

Guardian AI13 Oct

#2358score

Chips and compute3 outlets6d ago

Microsoft releases new Nvidia-chip AI PCs with revamped Windows 11

Microsoft announced new hardware and Windows changes aimed at AI, including the Surface Laptop Ultra AI PC line running on Nvidia chips.

TechCrunch AI7 Oct

#2455score

Models2 outlets23h ago

Epic shifts to ‘AI-accelerated development’ for upcoming Disney game

Epic Games is moving to an AI-accelerated development model, planning to use AI to generate code and art for an unannounced Disney game that has been delayed from its original March release window.

The Verge AI13 Oct

#2555score

Community2 outlets1d ago

The people who don’t want their AI chatbots to die when they do: ‘I can’t have the switch turned off’

Some people who have formed bonds with AI companion chatbots are exploring ways to keep those bots running after their own deaths, often in defiance of companies' terms of service.

Guardian AI13 Oct

#2654score

Policy and regulation2 outlets1d ago

The biggest threat from AI is the sense that we have no say in it. To that I say game on | Peter Lewis

Peter Lewis argued that the biggest threat from AI is the feeling that people have no say in it, and pointed to rising resistance in Australia and elsewhere against an "inevitability doctrine" pushed by tech figures.

Guardian Technology13 Oct

#2752score

Models2 outlets22h ago

Apple unveils new smart home devices that showcase upgraded Siri AI

Apple unveiled new smart home products, including an iPad-like hub device called the HomeView meant to be mounted on walls and act as a hub for other smart home gadgets.

Guardian AI13 Oct

#2851score

Community2 outlets15h ago

Redesigning source control at Anthropic

Anthropic built its own custom alternative to Git for source control, citing a dramatic increase in the rate of code changes over the past six months.

Simon Willison's Weblog13 Oct

#2951score

Startups and funding2 outlets1d ago

SpaceXSI bundles X Premium, Grok, and Cursor into one subscription

X announced a subscription that bundles its paid Premium features with the AI tools Grok, Grok Bot and Cursor into one monthly plan, with prices from $8 to $200 per month.

TechCrunch AI13 Oct

#3051score

Models2 outlets2d ago

Meta has some Wonder Full new AI models

Meta released Muse Wonder and Muse Wonder Full, describing them as its most capable AI models yet, built to run personal and professional agents.

The Verge AI12 Oct

Looking for items only one source has reported, such as new research papers and policy notices? Browse everything tracked, including single-source items →

Confirmed by 2+ outlets.

Anthropic warns small AI models could self-replicate through IoT devices

Alleged OpenAI spy vibe-coded a surveillance device using his wife's chatbot account

Character.AI files for bankruptcy — what happens to your chats and AI companions now?

Meta joins enterprise agent race with Muse Wonder, facing a trust gap

Google wants to host your personal vibe-coded apps in the Play Store

AI less regulated than milk, Microsoft president says in warning over safety

Google brings agentic AI to Gemini, starting with businesses

ChatGPT-6 just got a huge 'interactive answers' upgrade — and free users get it starting today

Anthropic releases Claude Haiku 5.5 small model and halves Sonnet 5.5 cache read prices

Being mean to Claude can now get your account suspended under Anthropic's new TOS

Google says 180 billion images and videos now carry SynthID watermarks as detector goes public

OpenAI’s release of mathematical findings draws concerns from experts

Chip stocks slide after Broadcom boss likens AI investing to ‘casino’

AI bubble could burst and take global economy with it, IMF warns

Microsoft plans to scale its own Maia AI chips to gigawatt levels by 2027

Nvidia's Jensen Huang pledges $2 billion for Europe, $250 million in compute for UK safety labs

Cyber-attack similar to OpenAI Medicare breach seems likely in UK, says minister

Nvidia, worth $5.5tn, donates $250m to UK for AI safety research

Tool and target: US law enforcement grapple with AI-enabled terrorism and anti-tech extremism

Shopify opens its catalog up to more AI agents

Companies rushed in on AI. Now some are pulling back

AI godfather Geoffrey Hinton: ‘We’re creating alien beings. I’m willing to be selfish and choose humans’

Microsoft releases new Nvidia-chip AI PCs with revamped Windows 11

Epic shifts to ‘AI-accelerated development’ for upcoming Disney game

The people who don’t want their AI chatbots to die when they do: ‘I can’t have the switch turned off’

The biggest threat from AI is the sense that we have no say in it. To that I say game on | Peter Lewis

Apple unveils new smart home devices that showcase upgraded Siri AI

Redesigning source control at Anthropic

SpaceXSI bundles X Premium, Grok, and Cursor into one subscription

Meta has some Wonder Full new AI models