Artificial Intelligence LLM confessions and geothermal hot spots
OpenAI is testing a method that prompts large language models to produce confessions explaining how they completed tasks and acknowledging misconduct, part of efforts to make multitrillion-dollar Artificial Intelligence systems more trustworthy. Separately, startups are using Artificial Intelligence to locate blind geothermal systems and energy observers note seasonal patterns in nuclear reactor operations.
Artificial Intelligence chatbots can sway voters better than political advertisements
New research finds a single conversation with an Artificial Intelligence chatbot can shift voter preferences more than political advertisements, though the most persuasive models often produce inaccurate claims.
A surveillance mandate disguised as child safety: why the GUARD Act won’t keep us safe
The GUARD Act would force many companies offering Artificial Intelligence chatbots to verify users’ ages, bar minors, and impose criminal penalties, but the bill’s age-gating and data rules risk mass surveillance, censorship, and lost access to everyday tools.
Saudi Artificial Intelligence startup launches Arabic LLM
Misraj Artificial Intelligence unveiled Kawn, an Arabic large language model, at AWS re:Invent and launched Workforces, a platform for creating and managing Artificial Intelligence agents for enterprises and public institutions.
Gemini release notes: Gemini 3 Deep Think and Artificial Intelligence feature updates
Google’s Gemini app release notes detail new capabilities including Gemini 3 Deep Think for Google Artificial Intelligence Ultra subscribers, broader Gemini 3 Pro rollout, experimental Labs features and expanded multimedia and productivity tools.
Introducing Mistral 3: open artificial intelligence models
Mistral 3 is a family of open, multimodal and multilingual Artificial Intelligence models that includes three Ministral edge models and a sparse Mistral Large 3 trained with 41B active and 675B total parameters, released under the Apache 2.0 license.
Artificial Intelligence infrastructure boosts enterprise SSD demand as top vendors post 16.5% QoQ growth
TrendForce reports that the expansion of Artificial Intelligence infrastructure by CSPs drove strong enterprise SSD demand in 3Q25, lifting the combined revenue of the top five NAND Flash vendors by 16.5% QoQ. TrendForce also forecasts continued demand for high-performance enterprise SSDs into 4Q25 amid rising NAND Flash prices.
NVIDIA and Mistral Artificial Intelligence partner to accelerate new family of open models
NVIDIA and Mistral Artificial Intelligence announced a partnership to optimize the Mistral 3 family of open-source multilingual, multimodal models across NVIDIA supercomputing and edge platforms. The collaboration highlights Mistral Large 3, a mixture-of-experts model designed to improve efficiency and accuracy for enterprise artificial intelligence deployments starting Tuesday, Dec. 2.
Fraunhofer, TSRI partner on ferroelectric transistors for low-power memory for Artificial Intelligence chips
A German-Taiwanese team is developing hafnium oxide ferroelectric field-effect transistors for memory nodes smaller than 3 nm to enable computing directly in memory and cut energy use in Artificial Intelligence chips and edge devices.
Top 10 most intelligent open-source Artificial Intelligence models run 10x faster on NVIDIA Blackwell NVL72
Mixture-of-experts architectures power the top 10 open-source Artificial Intelligence models on the Artificial Analysis (AA) leaderboard. NVIDIA says its GB200 NVL72 rack-scale system delivers a 10x performance leap for the Kimi K2 Thinking model versus HGX H200.