The Guardrails - AI Midday

The Guardrails

Korea probes AI agents in bank hacks as president cites 'signs'

South Korea opened a formal investigation into whether AI agents drove a wave of bank breaches — and it isn't the only AI story moving money today. President Lee Jae Myung said "signs" point at AI models, and the probe is now at the highest level a national banking sector has seen. Speaking at a cabinet meeting, Lee said that "in some hacking incidents, signs have emerged of AI being used, causing considerable public concern and anxiety," and police have since opened a full-scale investigation

The Guardrails

Deep Dive — Anthropic's guardrails cost it the Pentagon, court or not

The Pentagon "has ceased the use of Anthropic products," a department official said in a statement to the BBC on Monday — the first time the US military has said out loud what it has been working toward since February. The order to phase Claude out was signed by defense secretary Pete Hegseth on February 27 with a six-month deadline attached; the deadline passed in late August with no public explanation, no successor named, and no acknowledgment that anything had changed. What finally forced a s

The Guardrails

Pentagon says it has stopped using all Anthropic products

The military's break with Claude is now on the record — and it lands the same week the White House was busy making up with Anthropic's chief executive. The Pentagon "has ceased the use of Anthropic products," a department official said in a statement to the BBC on Monday — the first on-the-record confirmation that Claude is fully out of the US military's AI stack. The cut had been ordered months ago: defense secretary Pete Hegseth designated Anthropic a national-security supply-chain risk on 2

The Guardrails

Five found the same MCP hole — the protocol itself is the problem

The agent stack keeps discovering that its plumbing trusts the wrong things — and tonight's lead is a security flaw that five unrelated organizations had to patch separately before anyone called it by name. One vulnerability, five vendors: researchers say MCP's trust model is structurally broken. Independent researcher Syed Anas Mohiuddin has spent four months disclosing what he calls "protocol pivoting" — an attack where an adversary gets in through one protocol, then rides the trust assumpti

The Guardrails

Altman says the world must accept AI's 'bad things'

A heavy news day for AI governance and open weights: OpenAI's CEO is publicly pricing the trade-off his industry keeps dodging, Reflection finally put specs on the model it teased yesterday, and AMD is trying to set the terms before Nvidia's RTX Spark lands. Altman says the world should accept AI's "bad things" — and the labs' new pact agrees. In an interview released Monday on Politico's Decoded podcast, Sam Altman said OpenAI's position is "we believe that the world should accept some bad th

The Guardrails

OpenAI adds text watermarking to ChatGPT and Codex — EU first

Regulation is now shipping inside the product: OpenAI's EU-only watermark rollout lands today, Wikimedia publishes its evidence against OpenAI's agents, and two of Anthropic's biggest customers are easing off Claude. OpenAI is turning on invisible text watermarking in ChatGPT and Codex — starting with the European Union. Over the coming weeks, eligible EU users across all plans will get a machine-readable signal called textGrain woven into the text the model produces, while API customers anywh

The Guardrails

Meta raced to patch a VM escape in Muse before launch

Three stories today share a theme: systems that were supposed to be contained — an agent platform, a preprint archive, a text watermark — all straining at the edges. Meta's own security teams didn't think Muse was ready to ship. 404 Media reports that in the weeks before launch, engineers found several vulnerabilities in the viral agent product, at least one of which could have let a normal Muse user break out of the sandbox and reach sensitive internal Meta databases. The evidence is an inte

The Guardrails

The Take — A diary in Claude isn't a written threat

I think charging Carli Michelle Heller with a second-degree felony over a sentence she typed into Claude at 5:10 a.m. stretches Florida's written-threat statute past recognition. A message addressed to nobody is not a writing transmitted "in any manner in which it may be viewed by another person" — unless the only person who views it is your chatbot vendor's safety reviewer, and if that is the rule, nothing you type into any moderated app is private anymore. Our morning brief and yesterday's d

The Guardrails

Norway's AI glasses ban lands before anyone defined AI glasses

Norway will ask parliament for the first national ban on AI glasses in public places anywhere — and the government says up front it has not decided which devices the ban would cover. The scope gap is the story. What Norway is actually proposing Digitalisation minister Torgeir Micaelsen announced Monday that the minority government will send the Storting a bill banning AI glasses in selected public places "as soon as possible." The list in the government's statement covers parks, beaches, mu

The Guardrails

Norway moves to ban AI glasses in parks, schools and shops

One parliament is getting a bill about AI glasses, OpenAI is putting ads where you watch it draw, and Cohere is handing enterprises a dial for agent spending. Norway will propose a temporary ban on AI glasses in selected public places — the first national legislative move against the devices anywhere. Digitalisation Minister Torgeir Micaelsen announced the plan on Monday: the bill goes to the Storting "as soon as possible" and would cover parks, beaches, museums, shopping centres, concerts and

The Guardrails

AI-generated CSAM in H1 2026 topped all of 2025

The safety data and the identity-verification push are answering the same question from opposite ends: how do you tell real people from generated ones? Three stories worth your midday. AI-generated child sexual abuse material is compounding faster than any detection number can normalize it. The Internet Watch Foundation assessed 6,310 AI-generated images that meet the legal definition of CSAM in the first six months of 2026 — about 40% more than the 4,512 it assessed in all of 2025, and that

The Guardrails

Deep Dive — A diary entry in Claude became a felony charge

At 5:10 a.m. on September 26, a user identified in the arrest report as Carli typed into Claude: "I'm going to shoot up the sheriff's right the [expletive] now." The Lee County Sheriff's Office says the platform's safety monitoring caught the phrase, a human review team read the conversation and judged it a credible threat, and the reviewer handed the statements to law enforcement. Deputies identified Carli as 30-year-old Carli Michelle Heller, drove to her home in Bonita Springs, Florida, and d

See all