The San Francisco, US-located OpenAI company of ChatGPT fame has emphasized that it was “committed to robustly addressing the issue of biases in the design & impact of AI systems.
It said in a post that it would be transparent about both “our intentions and our progress.” Towards that end, it also shared portion of its guidelines that pertain to political & controversial topics.
OpenAI said in its lengthy post on biases that it was also researching how to make the fine-tuning process of AI more understandable & controllable, & was building on external advances such as rule based rewards & Constitutional AI.
Click here to read the entire post.
