# OpenAI Explores How AI Defenses Transfer Across Different Attack Types
OpenAI has shared research on the transfer of adversarial robustness between different types of perturbations, examining whether AI systems
OpenAI · May 6, 2026
# OpenAI Fellows Program Completes Second Class, Opens Applications for Summer 2019
OpenAI announced that its Fall 2018 Fellows cohort has successfully completed their six-month apprenticeship program, with participants tran
OpenAI · May 6, 2026
# OpenAI Scholars 2019 Program Concludes With Eight Final Projects
OpenAI announced via Twitter that its second cohort of OpenAI Scholars has successfully completed the program, with all eight participants p
OpenAI · May 6, 2026
# OpenAI Hosts First Robotics Symposium in 2019
OpenAI announced via Twitter that it held its inaugural Robotics Symposium on April 27, 2019, marking a significant milestone in the AI rese
OpenAI · May 6, 2026
# OpenAI Calls for Industry-Wide Cooperation to Prevent AI Safety Race to the Bottom
OpenAI has published a policy research paper warning that competitive pressures in the AI industry could create a dangerous "collective acti
OpenAI · May 6, 2026
# Microsoft Invests $1 Billion in OpenAI to Build Artificial General Intelligence
OpenAI announced a major partnership with Microsoft that includes a $1 billion investment aimed at developing artificial general intelligenc
OpenAI · May 6, 2026
# OpenAI Gives Employees Weekly "Learning Day" for Skill Development
OpenAI announced on social media that the company dedicates every Thursday as "Learning Day," allowing employees to step away from daily tas
OpenAI · May 6, 2026
OpenAI Completes Staged Release of GPT-2 Language Model
OpenAI has announced the full release of GPT-2, its 774 million parameter language model, marking the completion of a cautious, staged rollo
OpenAI · May 6, 2026
# OpenAI Develops New Method to Test AI Defense Against Unknown Attacks
OpenAI has announced a breakthrough in AI security testing with a new approach to evaluate how well neural networks can defend against adver
OpenAI · May 6, 2026
# AI Agents Develop Unexpected Strategies Playing Hide-and-Seek
OpenAI has revealed that artificial intelligence agents spontaneously developed complex tool-using behaviors while playing a simple game of
OpenAI · May 6, 2026
OpenAI Teaches GPT-2 to Match Human Preferences Through Feedback
OpenAI has announced successful fine-tuning of its 774-million parameter GPT-2 language model using direct human feedback, marking an import
OpenAI · May 6, 2026
OpenAI Scholars Program Opens Applications for 2020 Class
OpenAI announced via Twitter that applications are now open for the third cohort of the OpenAI Scholars program in 2020. The OpenAI Scholars
OpenAI · May 6, 2026
OpenAI Teaches Robot Hand to Solve Rubik's Cube
OpenAI has announced a breakthrough in robotics: a human-like robot hand that can solve a Rubik's Cube using neural networks trained entirel
OpenAI · May 6, 2026
# OpenAI Releases Full GPT-2 Model After Staged Rollout
OpenAI has released the complete version of GPT-2, its language model with 1.5 billion parameters, marking the end of a carefully planned st
OpenAI · May 6, 2026
# OpenAI Releases Safety Gym to Make AI Training Safer
OpenAI has announced the release of Safety Gym, a new toolkit designed to help researchers develop reinforcement learning agents that follow
OpenAI · May 6, 2026
# OpenAI Highlights New Benchmark for Safe AI Exploration
OpenAI has shared research on benchmarking safe exploration in deep reinforcement learning, addressing a critical challenge in AI developmen
OpenAI · May 6, 2026
# OpenAI Releases Procgen Benchmark to Test AI's Learning Abilities
OpenAI has announced the release of Procgen Benchmark, a new testing suite designed to measure how well reinforcement learning agents can ge
OpenAI · May 6, 2026
# AI Models Get Worse Before They Get Better, OpenAI Research Reveals
OpenAI has documented a puzzling phenomenon called "double descent" that challenges conventional wisdom about training artificial intelligen
OpenAI · May 6, 2026
# OpenAI Demonstrates Deep Reinforcement Learning Breakthrough with Dota 2
OpenAI announced a significant achievement in artificial intelligence by successfully applying large-scale deep reinforcement learning to Do
OpenAI · May 6, 2026
# OpenAI Highlights Scaling Laws for Neural Language Models
OpenAI recently shared insights about scaling laws for neural language models, drawing attention to fundamental principles that have shaped
OpenAI · May 6, 2026