General Cyber News via Ars Technica Risk Assessment

Users complain of new "sycophancy" streak where ChatGPT thinks everything is brilliant.
Publish Date: 4/21/2025
read more -->

Frustrated software developer believed AI-generated message came from human support rep.
Publish Date: 4/17/2025
read more -->

New o3 model appears "near-genius level," according to one doctor, but it still makes mistakes.
Publish Date: 4/16/2025
read more -->

Prompt injections are the Achilles' heel of AI assistants. Google offers a potential fix.
Publish Date: 4/16/2025
read more -->

Early info is unreliable, but the site has been mostly unavailable for hours.
Publish Date: 4/15/2025
read more -->

OpenAI's brand-new "GPT-4.1" has a funky name but reasonable performance for the price.
Publish Date: 4/14/2025
read more -->

New announcement seems aimed at appeasing Trump, signals company shift.
Publish Date: 4/14/2025
read more -->

Snapshotting and AI processing a screen every 3 seconds. What could possibly go wrong?
Publish Date: 4/11/2025
read more -->

Even weirder: Why would Google give so many the "Featured" stamp for trustworthiness?
Publish Date: 4/11/2025
read more -->

New Anthropic research shows AI models often fail to disclose reasoning shortcuts.
Publish Date: 4/10/2025
read more -->

Company didn't notice its chatbot was being abused for (at least) 4 months.
Publish Date: 4/9/2025
read more -->

Two-tiered "Claude Max" expands rate limits and offers traffic priority to subscribers.
Publish Date: 4/9/2025
read more -->

Disclosure of tactics, techniques, and procedures provides rare glimpse into secretive group.
Publish Date: 4/8/2025
read more -->

Carmack: "Power tools" like AI are "central to all the progress in computers."
Publish Date: 4/8/2025
read more -->

Touted 10M token context proves elusive, while early performance tests disappoint experts.
Publish Date: 4/7/2025
read more -->

Used by nation-states and crime groups, fast flux bypasses many common defenses.
Publish Date: 4/4/2025
read more -->

Yes, encryption/decryption occurs on end-user devices, but there's a catch.
Publish Date: 4/3/2025
read more -->

Automated AI bots seeking training data threaten Wikipedia project stability, foundation says.
Publish Date: 4/2/2025
read more -->

Model context protocol standardizes how AI uses data sources, supported by OpenAI and Anthropic.
Publish Date: 4/1/2025
read more -->

A safe and proper rewrite should take years not months.
Publish Date: 3/29/2025
read more -->