Imagine an AI system that can do anything better than you. That’s superintelligence, and it could be here sooner than you think.
In a blog post published on Tuesday, three prominent leaders of OpenAI, an AI research and deployment company, shared their views on the challenges and opportunities of developing superintelligence, a hypothetical AI system that surpasses human intelligence in all domains.
Sam Altman, Greg Brockman, and Ilya Sutskever, who are the CEO, CTO, and Chief Scientist of OpenAI respectively, argued that superintelligence could be created within the next ten years and that it would have unprecedented power and impact on humanity.
“In terms of both potential upsides and downsides, superintelligence will be more powerful than other technologies humanity has had to contend with in the past. We can have a dramatically more prosperous future; but we have to manage risk to get there. Given the possibility of existential risk, we can’t just be reactive,” they wrote.
To ensure that superintelligence is developed in a safe and beneficial manner, they proposed three ideas:
- Some degree of coordination among the leading development efforts to ensure that the growth of AI capability at the frontier is limited to a certain rate per year and that safety and social integration are maintained.
- An international authority that can inspect, audit, test, and regulate superintelligence efforts above a certain capability or resource threshold, similar to the IAEA for nuclear energy.
- The technical capability to align superintelligence with human values and goals, which is an open research question.
They also emphasized the importance of public oversight and democratic decision-making on the bounds and defaults for AI systems, as well as the need to allow innovation and experimentation with models below a significant capability threshold.
They concluded by stating their reasons for pursuing superintelligence: “First, we believe it’s going to lead to a much better world than what we can imagine today… Second, we believe it would be unintuitively risky and difficult to stop the creation of superintelligence.”
OpenAI was founded in 2015 by a group of prominent entrepreneurs and researchers, including Elon Musk and Sam Altman. It received a $1 billion investment from Microsoft in 2019 and a $10 billion investment in 2023. It is known for its groundbreaking AI systems such as GPT-4, ChatGPT, DALL·E, and Point-E.