OpenAI Cybersecurity Grant Program
Our goal is to facilitate the development of AI-powered cybersecurity capabilities for defenders through grants and other support.
Mozilla Ventures Invests in Fiddler, Fueling Better AI Trust
We’re excited to announce an investment by Mozilla Ventures to help us build transparency and trust in AI.
Improving mathematical reasoning with process supervision
We’ve trained a model to achieve a new state-of-the-art in mathematical problem solving by rewarding each correct step of reasoning (“process supervision”) instead of simply rewarding the correct final answer (“outcome supervision”). In addition to boo…
Thinking of AI as a Public Service
AI is often viewed through a binary positive or negative lens. Saad Ansari, Director of AI at Jasper AI, offers a novel view on AI as a public service.
Making Image Explanations Human-Centric: Decisions Beyond Heatmaps
Learn how explanations of Computer Vision model predictions can be made more human-centric when using integrated gradients to generate saliency maps.
Running Large Language Models Privately – privateGPT and Beyond
A discussion on data privacy and privacy-preserving machine learning for LLMs