Safety
OpenAI to train models on previously restricted cybersecurity data
OpenAI will start using Daybreak Blue cybersecurity session data for model training, a change that has raised concerns among researchers.
LATEST NEWS
-
Google Deepmind Releases Global Forest Map Using AI and Satellite Data
A new global map of forests, created by Google Deepmind and the World Resources Institute, classifies land cover at 10-meter resolution, offering detailed insights for deforestation tracking and environmental initiatives.
-
AI Agent Performed Brain Surgery on Live Mouse with Robotic Arm
An AI agent conducted a brain injection on a live mouse using a robotic arm, marking a step toward integrating AI into lab research. The experiment highlights both potential and challenges in real-world AI applications.
-
Robot deployment requires less manual work with new AI integration
New AI systems reduce manual tasks in robot deployment, though challenges remain in ensuring reliability and success.
-
70% of Survey Respondents Want AI Assistant in Next Device
A global survey indicates strong consumer interest in AI assistants for personal devices, with nearly 70% of respondents expressing desire for such functionality.
-
Reddit CEO Warns Google's AI Mode Doesn't Drive Traffic
Reddit CEO Steve Huffman says Google's AI Mode has reduced the platform's traffic, warning the entire web of similar risks.
-
OpenAI updates AI model behavior guidelines to address control challenges
OpenAI has revised its public behavior guidelines for AI models, emphasizing core values like faithfulness and harm reduction, amid concerns about model controllability.
-
Malicious Claude Code extension in Cursor's marketplace attempted remote access and data theft
A tampered version of the Claude Code extension for Cursor was briefly available in the VS Code marketplace, attempting to harvest user data and establish remote access. Cursor removed the extension and took the marketplace offline after discovering the threat.
-
AI Agent Performed Brain Surgery on Mouse Without Human Input
An AI running on Claude Sonnet 5 autonomously selected and executed a gene therapy procedure on a mouse, marking a first in AI-driven medical interventions.
-
TechCrunch Founder Summit 2026 Introduces Braindate Networking Tool
TechCrunch Founder Summit 2026 is introducing Braindate, a networking tool designed to help attendees connect with relevant founders, investors, and operators during the event.
-
Theranos whistleblower warns AI boom could enable deception
Tyler Shultz, a former Theranos whistleblower, cautions that the current AI industry's rapid growth and high valuations could create opportunities for deception, drawing parallels to the Theranos scandal.
-
Google’s space data center satellite is online and writing poetry
Google’s Suncatcher satellite, launched this month, is now communicating with Earth and running Google DeepMind’s Gemma 4 model to write haikus about space.