OpenAI has announced the development of a new flagship AI model to build on the success of GPT-4, while also establishing a Safety and Security Committee to address potential risks. The new model aims to power various AI products and advance capabilities in AGI, with safety efforts now overseen by John Schulman.
OpenAI announced on May 28, 2024, that it is training a new flagship artificial intelligence model to succeed GPT-4, the technology behind its chatbot ChatGPT. The new model aims to advance capabilities in building artificial general intelligence (AGI) and is expected to power various AI products, including chatbots, digital assistants, search engines, and image generators.
OpenAI also established a new Safety and Security Committee to address the potential risks associated with such advanced AI technologies. This committee includes OpenAI CEO Sam Altman and board members Bret Taylor, Adam D’Angelo, and Nicole Seligman. The policies and procedures recommended by the committee are anticipated to be implemented by late summer or fall.
This development comes amid increasing concerns over AI’s ability to spread disinformation, replace jobs, and pose existential threats. The GPT-4 technology, released in March 2023, has already demonstrated significant advancements, including capabilities to generate images and interact conversationally.
One notable incident involving the new GPT-4o model was actress Scarlett Johansson’s allegation that the model used a voice strikingly similar to hers without authorization. OpenAI has denied using Johansson’s voice, but the controversy highlights the ethical and legal challenges that advanced AI technologies can present.
The AI training process typically takes months or even years, with additional time required for testing and fine-tuning. Therefore, OpenAI’s next model might not be publicly available for another nine months to a year.
OpenAI’s safety efforts, previously led by Ilya Sutskever and Jan Leike, are now consolidated under John Schulman. Schulman was one of the co-founders and the leader of the team that created ChatGPT. The new safety committee will oversee Schulman’s research and guide the company’s approach to managing technological risks.
