Sam Altman admits AI models 'scare' him and understands why world is anxious

OpenAI CEO Sam Altman has admitted that the rapid rise of AI can be "scary", acknowledging the anxiety surrounding increasingly powerful AI models. Speaking on X (formerly Twitter), Altman shared his thoughts on the tension between excitement and fear as AI capabilities improve. The CEO emphasized that while the company has been living with this discordant feeling for some time, it is even more distressing for others outside the organization.

Altman stressed that understanding where AI is headed and how models perform in the real world is crucial, but more importantly, the world needs AI systems that are aligned with human values to navigate the future phases of this technological transition.

Altman believes that simply worrying about AI is not enough. He called for society, AI companies, and other stakeholders to take the situation seriously and evolve alongside the technology, prioritizing safety and alignment as top concerns. "We have been living with the tension between being excited and anxious about progress for some time, and it is still discordant for us," Altman stated.

"We know it is much more discordant for other people. And yet, we believe strongly that the world needs to understand where AI is going and how models perform in the real world. More importantly, we believe the world will need aligned AI to manage the future phases of this transition."

In the post, Altman expressed the company's pride in Astra's capabilities while also recognizing the need for caution. He acknowledged that while Astra is impressive, OpenAI believes that the current phase of development warrants a measured approach to ensure that safety standards are met as new capabilities are introduced. The CEO highlighted the company's recent focus on safety priorities, emphasizing that capabilities and safeguards must advance together.

OpenAI plans to launch its next model soon, signalling a continued commitment to responsible AI development.

Altman concluded by acknowledging the uncertainty surrounding AI's consequences and the need for a collaborative approach between society and technology. He urged the world to take AI developments seriously, emphasizing the importance of understanding where AI is headed and ensuring that AI systems are aligned with human values.

Altman expressed hope that the community continues to prioritize AI safety and believes that an iterative loop where society and technology evolve together will lead to the highest chance of success in managing the future of AI.