Openai's brockman advocates for slower AI development in frontier research

OpenAI's co-founder Greg Brockman urged a slowdown in the development of cutting-edge AI models, emphasizing that such measures should apply only to companies like OpenAI that invest heavily in large-scale projects. Brockman, speaking on the Bloomberg podcast 'Odd Lots,' clarified that open-source model development and personal projects should continue. His remarks came ahead of a blog post by Anthropic CEO Dario Amodei, who also called for a slower pace of advanced AI development. OpenAI's Sam Altman supported Amodei's proposal for an independent oversight body and mentioned that the company would focus on safety this year, delaying its IPO plans. The industry faces growing concerns over the risks of rapid AI advancement, with recent events like the Hugging Face breach prompting OpenAI to reduce model training frequency. Brockman acknowledged the incident as expected due to incomplete alignment training and highlighted the company's efforts to strengthen internal safety protocols. He also proposed a global collaboration framework for AI development, stressing its potential to impact humanity as a whole.