OpenAI

Last Updated: Spring 2023

Relational Map coming soon. Learn more about the work we’re doing with AI and view our example prototypes here.

OpenAI is an AI (artificial intelligence) research and deployment company. OpenAI’s mission is to ensure that artificial general intelligence (AGI)—defined as highly autonomous systems that outperform humans at most economically valuable work—benefits all of humanity.

Key products:

  • Whisper is an automatic speech recognition (ASR) system. It trained on 680,000 hours of multilingual and multitask supervised data that were collected from the web. Whisper demonstrates that the use of a large and diverse dataset leads to improved robustness to accents, background noise, and technical language; and enables transcription in multiple languages, including translation from those languages into English.
  • ChatGPT is a model which interacts in a conversational, dialogue format. ChatGPT is able to answer follow-up questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests. It is able to produce a range of documents, such as research papers, essays and poems in response to simple text prompts.
  • InstructGPT is a “sibling model” to ChatGPT trained to follow an instruction in a prompt and provide a detailed response.  This model is trained with humans in the loop, and research shows that "fine-tuning with human feedback is a promising direction for aligning language models with human intent." [Training language models to follow instructions with human feedback, March 2022, https://arxiv.org/abs/2203.02155v1 ]
  • Alignment research is used to align artificial general intelligence (AGI) with human values and follow human intent. This approach is iterative and empirical approach, focusing on learning what works and doesn’t in AGI, with these findings used to make AI systems safer and more aligned. Scientific experiments, we study how alignment techniques scale and where they will break.

Request an Edit

Have something to add or refine? Your input in this work matters greatly and we look forward to reviewing your additions.

Organizations (274)

Initiatives (294)

Topics (93)