OpenAI is taking significant steps to prioritize AI safety by establishing a dedicated Safety and Security Committee. This move comes as the company begins training a new AI model intended to replace the current GPT-4 system powering its widely used ChatGPT chatbot.
The newly formed committee will be responsible for providing recommendations to OpenAI’s board on essential safety and security matters related to the company’s projects and operations. In the first 90 days, the committee will assess and enhance OpenAI’s processes and safeguards. After this period, they will present their findings and suggestions to the board, followed by a public update on the adopted recommendations.
This initiative is set against a backdrop of internal debates on AI safety within OpenAI, highlighted by the resignation of researcher Jan Leike, who criticized the company for prioritizing product development over safety. Additionally, Ilya Sutskever, OpenAI’s co-founder and chief scientist, stepped down, leading to the dissolution of the \"superalignment\" team dedicated to addressing AI risks.
As Microsoft-backed OpenAI continues to develop advanced chatbots with generative AI capabilities, concerns surrounding the safety and ethical implications of increasingly powerful AI models have become more prominent. The company's latest effort includes the training of what they call the \"next frontier model,\" which is expected to propel them closer to achieving Artificial General Intelligence (AGI).
Reference(s):
OpenAI sets up safety committee as it starts training latest AI model
cgtn.com
