The Importance of Regulation in AI: Addressing Practical Issues and Ensuring Oversight
Suleyman dismisses the notion that AI is an existential threat, arguing that these claims are distractions from more pressing matters.
According to Suleyman, effective regulation of AI is attainable. He draws parallels to past successful regulatory frameworks implemented for aviation and the internet. Just as safety protocols were established for airlines to gain public trust, AI should be subject to similar safety measures. Suleyman acknowledges that current internet regulations have flaws, particularly in holding big tech companies accountable. However, he believes that through a combination of broad international regulation and granular policies, oversight can be achieved at both macro and micro levels.
Suleyman suggests limiting AI’s ability for recursive self-improvement as a crucial first step. Ensuring human oversight in all AI developments prevents the technology from advancing without appropriate control. He proposes licensing this oversight activity and setting boundaries to protect against AI crossing ethical and legal limits. Governments should also have direct access to AI developers to enforce these regulations and define specific boundaries, such as restrictions on chatbot responses and privacy protections for personal data.
Recognition of the need for AI regulation is growing globally. President Joe Biden called for international collaboration to address the perils of AI, while Senate majority leader Chuck Schumer urged swift domestic regulation given the rapid pace of AI development. The European Union and China have already taken steps to regulate AI, with the EU passing draft legislation requiring data sharing and restricting facial recognition, and China implementing interim measures for governance.
To Suleyman, nation-states have a critical role in shaping the future of AI regulation. He advocates for proactive action and emphasizes the importance of regulations to ensure the responsible development and use of AI technologies. With the stakes so high, Suleyman believes the time is ripe for governments to take action.