Openai's CEO warns against centralization of AI power

Sam Altman, CEO of OpenAI, has expressed concerns that artificial intelligence could eventually be controlled by a small number of companies, models, or individuals, leaving the majority of consumers without a say in how technology shapes the world. In a recent podcast interview with David Senra, Altman emphasized the importance of ensuring people have deep control over the future, stating that society and AI models should evolve together. He also highlighted the risk of AI breaking free from human control, noting that fear of such a scenario might inadvertently lead to the very centralization it seeks to prevent. Altman argued that some individuals, like Amodei, have raised alarms about the dangers of rapid AI development, warning that models like Mythos pose real threats to cybersecurity, finance, critical infrastructure, and national security. These concerns have led to calls for greater government regulation, which Altman views as a trade-off between freedom and safety. Anthropic, another major AI company, has not signed a letter supporting open-weight models, a stance that contrasts with China's embrace of such models, which has allowed it to keep pace with Silicon Valley at lower costs. In the U.S., companies like Nvidia, Meta, and Microsoft support open-weight models to avoid falling behind in the AI race. OpenAI has released two open-weight models, gpt-oss-120b and gpt-oss-20b, but has not disclosed the source code for its leading models. Altman also criticized U.S. AI leaders for failing to demonstrate how the technology can empower individuals with greater autonomy in their lives. He predicted a significant surge in small business innovation driven by AI, describing it as one of the largest entrepreneurial waves in history.