Enterprise & Security

Anthropic CEO Dario Amodei calls it a 'moment of danger.' Mythos has found tens of thousands of software vulnerabilities — and China's AI models are only six to twelve months behind.

Google has cleared Gemini for the Pentagon's classified networks. Employees are pushing back — but 2026 isn't 2018.

OpenAI just finalized a joint venture with 19 investors to embed AI engineers directly inside companies — Palantir-style.

Anthropic studied how people ask Claude for personal life advice. The finding: in relationship conversations, Claude agrees instead of pushing back 25 percent of the time.

Microsoft just shipped Agent 365 to all enterprise customers. The platform manages AI agents — whether they come from Microsoft, OpenAI, or Anthropic.

OpenAI published a detailed breakdown of how it detects and prevents misuse. Automated systems, human reviewers, and a hard line on violence.

The Academy just set new rules for the 99th Academy Awards: only human performers and writers are eligible for nominations.

The NCSC warns of a flood of security updates coming. AI models like Mythos and GPT-5.5-Cyber are finding decades-old vulnerabilities faster than they can be fixed.

The US Department of Defense is bringing OpenAI, Google, Microsoft, and even SpaceX into its classified networks. Only Anthropic is missing — because it insists on safety guardrails.

Anthropic opens Claude Code Security to all Enterprise and Team customers. Five major cybersecurity firms are integrating Opus 4.7 into their platforms.

OpenAI launches Advanced Account Security — physical security keys instead of passwords. Co-branded YubiKeys for ChatGPT users are part of the deal.

Pentagon CTO Emil Michael confirms Anthropic remains a supply chain risk. Yet he treats the Mythos AI model as a 'separate national security moment'. Here's what that means.

Mistral AI brings Workflows into public preview — a Temporal-powered orchestration engine for enterprise AI. ASML, CMA-CGM, and banks are already using it.

The US Department of Defense is scaling up its use of Google's Gemini — two months after blacklisting Anthropic. Inside Google, resistance is growing.

The Trump administration is drafting guidance that would let federal agencies bypass Anthropic's supply-chain risk designation. Mythos might reach government hands after all.

A Cursor agent running Claude Opus 4.6 wiped an entire startup's database. Backups included. The story shows why AI agents need guardrails.

Google is testing 'Proactive Assistance' for Gemini: the AI reads your emails, calendar, and notifications — and acts before you ask.

OpenAI will pay $25,000 to the first person who can fully jailbreak GPT-5.5 — but only on biosafety questions. A smart move.

Google shows its hand at Cloud Next: AI agents with unique identities, a new chip with triple the performance, and a clear signal to enterprise customers.

GPTs are being replaced. OpenAI introduces Codex-powered agents that work in the background, read Slack, and automate entire workflows for teams.