15 articles
OpenAI provides a shared playbook for conducting trustworthy third-party evaluations of frontier AI systems.
OpenAI presents a comprehensive Frontier Governance Framework that addresses AI safety, security, and risk management in alignment with evolving global regulations.
OpenAI introduces content provenance technologies to identify and verify AI-generated media for greater transparency and trust.
This article discusses new ChatGPT safety updates that enhance context awareness in sensitive conversations.
OpenAI has introduced Trusted Contact, a new safety feature in ChatGPT that enables adults to designate a trusted person who may receive notifications if the AI system detects serious safety concerns related to self-harm.
OpenAI introduces a European Youth Safety Blueprint and €500,000 grant program to support youth safety and wellbeing with AI.
OpenAI describes its approach to protecting community safety in ChatGPT through multiple layers of protection.
GPT-5.5 is a new AI model from OpenAI designed to handle complex, real-world tasks with improved autonomy and tool usage.
The GPT-5.5 Bio Bug Bounty is a red-teaming initiative designed to identify security vulnerabilities in AI systems related to biological safety.
OpenAI introduces a Safety Bug Bounty program focused on identifying AI-specific abuse and safety risks.
OpenAI introduces prompt-based teen safety policies to help developers build safer AI experiences for younger users.
This post outlines the safety-first approach taken in building Sora 2 and the Sora app as a video generation model and social creation platform.
OpenAI explains how it monitors internal coding agents for misalignment using chain-of-thought analysis.
OpenAI Japan has announced the Japan Teen Safety Blueprint, a framework designed to protect teenagers using generative AI.
