OpenAI CEO Sam Altman announced that the company has attained a fundamental understanding of artificial general intelligence (AGI) and is now focusing on the development of superintelligence, with the expectation that AI agents may join the workforce by 2025.
OpenAI aims for superintelligence by 2025
In a blog post reflecting on the past nine years since OpenAI’s inception as a non-profit research lab, Altman noted that the company became notable in the tech industry after launching ChatGPT in November 2022. He indicated that the initial name for the AI chatbot was “Chat With GPT-3.5.”
Altman asserted that OpenAI has mastered the traditional methods of building AGI, stating, “We continue to believe that iteratively putting great tools in the hands of people leads to great, broadly-distributed outcomes.” He elaborated that the introduction of superintelligent tools could dramatically accelerate scientific discovery and innovation, ultimately enhancing human capabilities and increasing global prosperity.
“We are beginning to turn our aim beyond that, to superintelligence in the true sense of the word. We love our current products, but we are here for the glorious future. With superintelligence, we can do anything else. Superintelligent tools could massively accelerate scientific discovery and innovation well beyond what we are capable of doing on our own, and in turn massively increase abundance and prosperity.”
-Sam Altman
He expressed confidence in achieving superintelligence in the coming years, emphasizing the need for careful actions while maximizing broad benefit and empowerment.
Rumors suggest that OpenAI is working on its next model, GPT-5, codenamed project Orion. The company is also expected to release the o3 series of AI models, which focus on advanced reasoning, later this year. Altman acknowledged that the o3 model has performed exceptionally well on benchmarks, scoring almost 90% on the ARC-AGI benchmark, surpassing human performance.
Is OpenAI’s promised copyright protection stalled for good?
OpenAI and Microsoft have collaborated to define AGI, with OpenAI stating that it can only achieve AGI by creating a system capable of generating $100 billion in profits. It was noted that superintelligence might seem speculative, but Altman is confident its significance will soon be apparent.
Furthermore, Altman remarked on the challenges of AI safety, stating that the transition to a landscape with superintelligence is not guaranteed and that OpenAI lacks definitive solutions for controlling a potentially superintelligent AI. He recognized that, as AI evolves, the organization must continue to focus on safety and alignment research. Despite some internal restructuring and staff departures related to safety concerns, Altman defended OpenAI’s track record on safety.
Featured image credit: Kerem Gülen/Midjourney
