OpenAI Releases GPT-5.6-Cyber to Defenders Amid Autonomous Attack Concerns

Axios · first seen 11.08.2026

OpenAI is introducing a cyber-permissive version of its GPT-5.6 model to vetted defenders to help prepare for autonomous cyberattacks, following safety delays with its Astra model.

Sources

Axios — OpenAI introduces a new cyber model amid fears of AI cyberattacks

Also in the feed that day

Tech&Science

Google Begins Hosting Rival App Stores in Play Store Following Epic Games Lawsuit

Ars Technica · TechCrunch

Tech&Science

July Recorded as the Hottest Month in Contiguous U.S. History

Axios

Tech&Science

England on Track to Eliminate Hepatitis C Ahead of Global Target

Hacker News

Tech&Science

Social Media Giants Face Continued Litigation After Failed Appeals in Addiction Lawsuits

TechCrunch

Tech&Science

Zoom Patches 'Zoomsday' Vulnerability Discovered via AI Prompts

The Verge

Tech&Science

Amazon Backs Fossil-Fuel Power Plant for Off-Grid AI Data Center

Ars Technica

Tech&Science

Scientists Warn 'Super' El Niño Could Push Amazon Rainforest Past Tipping Point

Nature

Tech&Science

OpenAI Launches New Cyber-Trained AI Model and Expands Daybreak Program

CNBC · TechCrunch

Tech&Science

New Ebola Vaccine Trial Launches in DR Congo Amid Outbreak

France 24

Tech&Science

Researchers Uncover SIM Card Vulnerabilities Allowing Remote Hijacking

The Register

Tech&Science

US Federal Agencies Warn of Gunra Ransomware Targeting Critical Infrastructure

The Register

Tech&Science

Nature Study Warns AI-Driven Oil Production Outweighs Renewable Benefits

Axios

Tech&Science

AI Tool Discovers Now-Patched Zoom Screen-Sharing Hijacking Bug

Wired

Tech&Science

Spotify to Label and Exclude 'AI Personas' From Recommendations

TechCrunch · The Verge

Tech&Science

Mark Zuckerberg Outlines Meta's Open AI Strategy and Funding Plans

Ars Technica · Business Insider · TechCrunch

Tech&Science

Amazon Data Center Projected to Become Top Source of US Emissions

Semafor

Tech&Science

Meta Releases Muse Glimmer, a 30-Billion Parameter Open-Weights LLM

The Register

Tech&Science

Anthropic Pledges to Implement Watermarking to Comply with EU AI Rules

The Register

Tech&Science

YouTube Doubles Watch Hour Requirement for Creator Monetization

Engadget · TechCrunch

Tech&Science

Anthropic to Implement Invisible Watermarks on Claude-Generated Text and Images

TechCrunch · The Verge

Tech&Science

Mosquito Populations Drop Across US as West Nile Virus Hits 22-Year High

The Independent

Tech&Science

AI Productivity Gains in Fossil Fuel Sector Could Increase Emissions, Study Finds

Wired

Tech&Science

Chinese Long March Rocket Explodes Shortly After Liftoff

Semafor

Tech&Science

LexisNexis Pulls Services Offline Following Suspicious Server Activity

The Register

Tech&Science

Nurses Push Back Against Clinical AI as CMS and FDA Propose RAPID Pathway

STAT News

Tech&Science

Meta Faces Legal Setbacks Over Child Safety as Google AI Executives Depart

The Guardian

Tech&Science

New Surveillance Tool 'SignalTrace' Tracks Individuals via Personal Devices

The Independent

Tech&Science

Samsung to Add FDA-Approved Hearing Aid Feature to Galaxy Buds

Engadget

Tech&Science

Mark Zuckerberg Announces $1 Billion Fund for Data Center Communities

Semafor

Tech&Science

Japan Launches H3 Rocket to Near-Complete Sovereign Satnav Constellation

The Register

Tech&Science

Mozilla Revokes Firefox Signing Key After GitHub Leak

The Register

Tech&Science

Spotify Introduces AI Persona Labels to Identify Non-Real Artists

Engadget

Tech&Science

Doctors Warn of CKM Syndrome Affecting Majority of Americans

The Independent

Tech&Science

NASA+ Streaming Service Expands to discovery+ and HBO Max

NASA