> ## Content Index
> Fetch the complete content index at: https://nordisk-ai.ghost.io/llms.txt
> Use this file to discover other available public pages before exploring further.

# OpenAI Strengthens Safety and Security Measures With New Committee
- URL: https://nordisk-ai.ghost.io/openai-strengthens-safety-and-security-measures-with-new-committee/
- Published: 2024-09-28T07:49:41.000Z
- Updated: 2026-08-27T17:21:24.000Z
- Author: Anders Hagen
- Tags: Cybersecurity, AI Risks, AI Security, Generative AI, OpenAI, Sweden, Sweden, #Migrated-1787841634512, #wp, #wp-post, #Import 2026-08-27 14:41, #Import 2026-09-07 13:41

OpenAI has implemented several new safety measures to manage the risks around developing and releasing its AI models. This comes through a newly established independent committee, the Safety and Security Committee (SSC), which will oversee safety and security processes.

### Independent Oversight for Greater Safety

The SSC will be an independent board committee led by Zico Kolter, head of the machine learning department at Carnegie Mellon University. The committee also includes Adam D'Angelo, co-founder and CEO of Quora, retired US General Paul Nakasone, and Nicole Seligman, former general counsel at Sony Corporation. They will review safety evaluations ahead of every major model release and have the authority to delay launches until safety risks have been addressed.

### Strengthened Security Measures

Cybersecurity is a central part of AI safety, and OpenAI is working to strengthen the protection of its research and products. The company is expanding its internal security team, improving information segmentation, and investing in new security initiatives. As part of this, it's considering creating an Information Sharing and Analysis Center (ISAC) for the AI industry, which would enable sharing of threat intelligence among players in the sector.

### Transparency in Safety Work

OpenAI aims to be transparent about its safety work. It has long published so-called "system cards" that describe its models' capabilities and risks. For example, the system cards for GPT-4o and the o1-preview have provided detailed insight into the safety work behind each launch, including results from external testing and risk assessments under the company's Preparedness Framework.

### Collaboration With External Organizations

OpenAI is expanding its partnerships with external security organizations and research institutions for independent review of its models. It has already begun collaborating with Los Alamos National Labs to study how AI can be used safely in scientific settings. In addition, OpenAI is working with AI safety institutes in the US and UK to study emerging safety risks and develop standards for trustworthy AI.

### A Unified Safety Framework for Model Development

OpenAI is building an integrated safety and security framework for model development and monitoring. This system will be based on risk assessments and require approval from the SSC. The goal is to ensure a consistent standard for model releases and to adapt safety strategies as models become increasingly advanced. OpenAI has also reorganized its research, safety, and policy teams to improve collaboration and ensure more effective integration of safety measures.