OpenAI co-founder emphasizes the need for AI labs to conduct safety tests on competing models to ensure responsible development and mitigate risks associated with advanced AI technologies. He advocates for a collaborative approach among AI developers to enhance safety standards across the industry.