All news
12055
Bond, a new social network that uses AI to wean you off doomscrolling
Startup Bond built a platform where AI actively nudges users to close the app and return to the real world, instead of keeping them glued to the screen.

Latitude unveils Voyage, a platform for building AI RPGs with living NPCs
Latitude, the maker of AI Dungeon, has announced Voyage, a platform for building full RPG worlds with characters powered by generative AI.

Clarifai destroys 3 million OkCupid user photos handed over for AI training, at the FTC's demand
In 2014, dating platform OkCupid secretly passed millions of user photos to Clarifai for facial recognition training. Under a settlement with the FTC, the images…

OpenAI details Sora 2 safety approach, from deepfakes to content labeling

ChatGPT launches visual shopping built on the Agentic Commerce Protocol

OpenAI releases prompt-based teen safety policies for gpt-oss-safeguard developers

OpenAI launches Safety Bug Bounty targeting vulnerabilities in agentic AI systems

OpenAI explains how the Model Spec works, the code of conduct for its AI models
OpenAI has openly described the principles that govern how its models behave, from a hierarchy of priorities to the split between operator and user permissions.

OpenAI raises $122 billion to expand frontier AI and compute capacity
OpenAI has closed a record $122 billion round. The money will fund next-generation compute, global expansion and growth of its enterprise AI business built on…

OpenAI launches a fellowship program for independent AI safety researchers

OpenAI publishes a plan to protect children from AI risks

OpenAI Announces a New Phase of Enterprise AI: ChatGPT Enterprise and Company-Wide Agents

CyberAgent speeds up its business with ChatGPT Enterprise and Codex

OpenAI expands Trusted Access for Cyber and launches GPT-5.4-Cyber for verified defenders
OpenAI is opening GPT-5.4-Cyber, a specialized model, to vetted cybersecurity professionals while tightening safeguards against misuse.

OpenAI updates Agents SDK with built-in sandbox and native harness for long-running agents
OpenAI has added native sandboxed execution and a model-native harness to the Agents SDK, making it easier to build safe, long-running agents.









