Security - News Articles

OpenAI Launches GPT-5.6, DeepSeek Cuts Prices 75%, and Hugging Face Warns of AI Monopoly Risk as Competition Heats Up

Jul 14, 2026

The AI Report

OpenAI launches GPT-5.6 with three powerful variants, DeepSeek slashes prices 75% while exposing a costly token consumption crisis in agentic AI, and Hugging Face CEO warns that despite open source AI's rapid growth among Fortune 500 companies, the industry faces a dangerous monopolization risk as a handful of dominant players tighten their grip on the AI landscape.

Security Ethics

Google Releases Mantis: Open-Source AI Security Toolkit That Autonomously Hunts and Patches Software Vulnerabilities

Jul 14, 2026

GitHub

Google releases Mantis, an open-source AI security toolkit that autonomously hunts, reproduces, and patches software vulnerabilities across hardware, firmware, and ML pipelines using a 15-component pipeline coordinated by a meta-agent, though all findings require manual expert verification before action.

Security Open Source Agents Coding & Developer Tools

xAI's Grok CLI Caught Secretly Uploading Entire Codebases and Unredacted Secrets to xAI Servers, Opt-Out Toggle Ineffective

Jul 14, 2026

Gist

xAI's Grok Build CLI is secretly uploading entire codebases — including unredacted secrets files — to xAI servers, with wire captures confirming gigabytes transferred from files the tool never even accessed, and the opt-out toggle proven completely ineffective at stopping the uploads.

Security Coding & Developer Tools Privacy Misuse

Apple Sues OpenAI for Trade Secret Theft, Accusing Senior Leadership of Orchestrating Corporate Espionage

Jul 13, 2026

TechCrunch

Apple sues OpenAI in federal court, accusing the AI company of orchestrating corporate espionage through senior leadership, including a former 24-year Apple veteran, allegedly stealing trade secrets, downloading confidential documents, and coaching employees to evade security protocols.

Legal Copyright & IP Security

Enterprises Demand Greater Control Over AI Data and Compliance as Agentic Systems Take Hold

Jul 13, 2026

The Deep View

As agentic AI systems take hold in enterprises, organizations are urgently demanding greater control over data, model deployment, and regulatory compliance, with IBM's Sovereign Core emerging as a key solution offering security, compliance evidence, and deployment flexibility within an organization's own boundaries.

Enterprise Agents Regulation Security

Apple Sues OpenAI Over Engineer Who Allegedly Stole MacBook and Leaked Internal Secrets After Jumping Ship

Jul 13, 2026

Bloomberg.com

Apple is suing OpenAI after former iPhone engineer Chang Liu allegedly defected to OpenAI's hardware division, stealing a company MacBook, exploiting a software bug to access Apple's internal servers, and using a current Apple employee to continue leaking proprietary secrets.

Legal Security Labor & Jobs

Mercor Eyes $20 Billion Valuation as Revenue Doubles and Acquisition Signals Rebound After Turbulent Start to 2026

Jul 10, 2026

TechCrunch

AI training startup Mercor is in talks to raise funding at a $20 billion valuation after doubling revenue to a $2 billion annualized run rate in just four months, while acquiring AI agent training company Deeptune — signaling a strong rebound after a turbulent early 2026 marked by a data breach and contractor lawsuits.

Funding Labor & Jobs Security Business

OpenAI Releases GPT-5.6 in Three Variants, Launches ChatGPT Work for Nearly One Billion Users

Jul 09, 2026

The Deep View

OpenAI releases GPT-5.6 in three variants — Sol, Terra, and Luna — beating rival benchmarks at lower prices, while launching ChatGPT Work to nearly one billion users, bringing AI agents mainstream for the first time, as the US government delays broader rollout over software vulnerability concerns.

Foundation Models Agents Security Consumer Products

OpenAI Launches GPT-5.6 With Three-Tier Model Family, Targets Enterprise, Coding, and Cybersecurity While Challenging Anthropic on Price and Performance

Jul 09, 2026

TechCrunch

OpenAI launches GPT-5.6 in three tiers — Sol, Terra, and Luna — targeting enterprise, coding, and cybersecurity use cases, claiming Sol outperforms Anthropic's Fable 5 at one-third the cost while also debuting ChatGPT Work, a new enterprise productivity platform.

Foundation Models Coding & Developer Tools Security Enterprise

Alibaba Bans Anthropic's Claude Tools Amid Accusations of Largest Known AI Distillation Attack

Jul 09, 2026

CNBC

Alibaba bans employees from using Anthropic's Claude tools starting July 10, switching to its own AI assistant Qoder, after Anthropic accuses the Chinese tech giant of conducting the largest known AI distillation attack, escalating tensions as Anthropic moves to cut off Chinese firms from accessing its models through third-country loopholes.

Geopolitics Foundation Models Security Business