OpenAI CEO Outlines AI Future: Ensuring Global Sharing of AGI Benefits

Sam Altman, CEO of OpenAI, recently published an article on his personal blog discussing how the company aims to enable "every person on Earth" to leverage Artificial General Intelligence (AGI) to achieve their goals and expand creativity.

Altman mentioned that part of this initiative includes some "odd-sounding ideas," such as providing everyone with a "compute budget" to ensure the benefits of AGI are widely distributed.

In the article, he wrote: "Historical impacts of technological progress suggest that most metrics we care about—such as health and economic prosperity—improve on average and in the long term. However, improvements in equality do not seem to be driven by technology alone; success in this area may require new approaches. In particular, the balance of power between capital and labor could easily be disrupted, which might necessitate early intervention."

Altman defines AGI as "a system capable of solving increasingly complex problems across multiple domains at a human level." He claims humanity is nearing the creation of such systems, a statement that could raise concerns that AGI might eventually lead to mass unemployment in many industries. Altman acknowledges that AGI's emergence will require "significant human oversight and guidance" to prevent such disruptions.

However, he insists that AGI systems "won't generate the most innovative ideas" and "excel in some areas but perform surprisingly poorly in others."

The true value of AGI will only be realized when these systems are run at massive scales. Altman envisions millions of super-large AI systems addressing "various issues in knowledge work."

Altman admits that such systems won't come cheap. He points out that the progress made in the AI industry so far shows that "spending arbitrary amounts of money can yield continuous and predictable performance improvements."

This explains why OpenAI is reportedly negotiating a funding round worth up to $40 billion, just three months after securing $6.6 billion in financing. The company is also collaborating with President Trump's "Stargate Project," partnering with Oracle and others to commit up to $500 billion in data center investments.

Nevertheless, Altman argues that despite the high costs of AI development, its usage costs decrease approximately tenfold every 12 months. Thus, while building more powerful AI systems requires significant investment, the technology will become increasingly accessible to everyone.

The rise of Chinese AI startups like DeepSeek seems to support this view, indicating that developing and training advanced AI systems may become more affordable. However, Altman maintains that achieving AGI will still require substantial financial investment.

Altman stated that if OpenAI succeeds in creating AGI-level systems, the company will undoubtedly need to make some "major decisions" and implement certain "restrictions related to safety," which might be unpopular.

Previously, OpenAI committed to ceasing competition with projects aligned in values and focused on safety as they approach true AGI development, instead opting to assist these projects to ensure safe AGI development. This commitment was made when OpenAI was still committed to its non-profit status. Now, the company is reorganizing itself as a for-profit entity, aiming for $100 billion in annual revenue by 2029.

Therefore, Altman emphasized that as OpenAI progresses toward AGI, the company's mission now focuses more on "empowering individuals" while striving to prevent such systems from being used by authoritarian governments for "mass surveillance to control populations" and to avoid "loss of autonomy."

This approach might mean OpenAI needs to become more transparent about its AI systems. Recently, Altman admitted that the company has "been on the wrong side of history regarding open-source practices," maintaining strict confidentiality over the codebases and training data of its most powerful AI systems.

Altman said: "Many of us anticipate that people will need greater control over technology than ever before, including more openness, while accepting the trade-offs between safety and individual empowerment."