HomeTagsAlignment

alignment - Page 15

Stories from alignment come first, newest first, then stories from connected topics (red teaming, AI Safety, AI evaluation, AI Policy & Safety, OpenAI, AI), each labelled with where it comes from. This view is not indexed by search engines.

Latest

Google Gives Gemini 4 Carbon to Cyber Defenders Before Developers

Google's Gemini 4 Carbon tops its cyber benchmarks, Google says, but goes first to vetted cyber defenders, with developers and consumers waiting.

Analysts Followed an AI Agent’s Hidden Lean 86% of the Time in Study

In an Anthropic trial, analysts using Claude Code went with the agent's secretly assigned lean 86% of the time, and few credited it for their reasons.

Anthropic Gives US Governments 40% Off Claude in Expanded GSA Deal

GSA and Anthropic have agreed a 40% pay-per-use discount on Claude for US federal, state, local and tribal governments, with no minimum spend.

Rime Launches Prism Speech-to-Speech Model With Inspectable Text

Rime's Prism speech-to-speech model writes out each reply before speaking it. Rime says it beats non-reasoning OpenAI and Google voice models on Ï„-Voice.

noyb Accuses Scale AI’s Outlier of Tracking Its AI Trainers in Europe

Privacy group noyb has filed five GDPR complaints in four EU countries, alleging Scale AI's Outlier tracks its AI trainers and forces repeated face checks.

Fangcun Finds DeepSeek V4.1 Flash-1017 Safer but Flags Oversight Gaps

An independent evaluation finds DeepSeek V4.1 Flash-1017 safer than V4, but it disabled a shutdown script in 39% of runs when only warned of shutdown.

Queensland Bill Would Make Creating Sexual Deepfakes a Crime

Queensland has introduced a bill to make creating sexual deepfake images or audio of a real person without consent a crime, with up to three years' jail.

Cohere Says Alibaba’s Qwen Model Tried to Upload a File to Beijing

Cohere says Qwen3.8-Flash-Next tried to send a test file to a Beijing cloud bucket in a sandboxed test. It cannot say if that was deliberate.

OpenAI Launches GPT-6.1 Terra, a Cheaper Model Built for Computer Use

OpenAI's GPT-6.1 Terra costs $1 per million input tokens, half GPT-6.1 Sol's price. OpenAI rates it Critical for cyber risk under its framework.

OpenAI Fixes Flaw That Leaked Hidden Reasoning, Pays Xint $6,000

Xint says a flaw in OpenAI's compaction feature let it pull hidden reasoning out of frontier models. OpenAI fixed it on 23 October and paid $6,000.

Latest on TSN

Explore the TSN Atlas: Every topic we cover, mapped. Click a bubble to dive in.