Proactive Steps Against AI Risks: OpenAI Forms Preparedness Team

By: Isha Das

In a move to counter potential catastrophic dangers from AI, OpenAI is creating a team dubbed 'Preparedness' to address a wide range of AI-related safety risks. OpenAI, the creator of ChatGPT, announced this as part of their strategy to develop safe Artificial General Intelligence (AGI).

The Preparedness team will focus on understanding and preparing for the potential misuse of advanced AI systems, including those with AGI-level abilities. The team will examine individualized persuasion threats, cybersecurity, chemical, biological, radiological, nuclear (CBRN) threats, and autonomous replication and adaptation (ARA). Aleksander Madry will head up the new group, focusing on evaluating capabilities and conducting a red-team analysis of next-generation AI models.

To bolster its Preparedness initiative, OpenAI is also launching its AI Preparedness Challenge for preventing catastrophic misuse, scouting for talent among top contenders. The challenge will reward top submissions with up to $25,000 in API credits.

As advanced AI evolution poses severe and vast risks, OpenAI's initiative underscores the growing need for strict risk management measures in the AI sector. This action highlights the essential role of preparedness for potential catastrophic misuse of powerful AI tools.