Superintelligence Must Follow Nuclear Energy Regulation

OpenAI published a blog post calling for efforts to regulate superintelligence similar to how nuclear energy is regulated. It suggests the eventual creation of an agency like that of the IAEA.

OpenAI, the creator of AI sensation ChatGPT, has proposed AI regulation, particularly AI superintelligence, akin to nuclear energy. The company published a post on May 22 about how AI’s rapid development could impact the world and the accompanying need for governance.

OpenAI Proposes Superintelligence Regulation

The post began by stating that “AI systems will exceed expert skill level in most domains, and carry out as much productive activity as one of today’s largest corporations.” Therefore, it is necessary to mitigate the risks of existential threat in the same manner as actions taken for nuclear energy.

The authors of the post—Sam Altman, Greg Brockman, and Ilya Sutskever—said that there would eventually be a need for something like The International Atomic Energy Agency (IAEA). This agency would help govern superintelligence efforts and would need a good deal of cooperation and coordination globally.

The post also clarified that it is important to allow “companies and open-source projects to develop models below a significant capability threshold, without the kind of regulation.”

OpenAI Considering Democratic Decisions on AI

OpenAI’s strategy to encourage the regulation of AI includes collective decision-making, similar to Wikipedia entries. President Greg Brockman announced that the company would assess the methods to gather input from various parties on decisions. The aim is to ensure a democratic process for decision-making.

Brockman said, “We’re not just sitting in Silicon Valley thinking we can write these rules for everyone,” though there will certainly be challenges in creating such a system. Like Sam Altman, he too was strongly focused on international cooperation for AI regulation.

The Race To Regulate AI Heats Up

The question of how to regulate AI is on everyone’s minds. Those in the thick of it are themselves calling for regulation. Altman has declined the opportunity to become America’s chief AI regulator. Instead, during a congressional hearing, he urged the U.S. to take responsibility for regulation.

The U.S. is likely to form an agency dedicated to AI oversight. Other countries are following suit, with China being particularly gung-ho about AI bans and limitations. Similar to the crypto market in the past, the AI industry may also experience abrupt and stringent regulation. This regulation would primarily concentrate on aspects such as privacy, preventing election interference, and establishing global standards.