OpenAI Recruits Zico Kolter A Professor At Carnegie Mellon University To The Board Of Directors

Zico Kolter (photo: Carnegie Mellon University)

JAKARTA - OpenAI officially recruited a Professor and Director of the AI-security-focused Machine Learning Department at Carnegie Mellon University, Zico Kolter as OpenAI's Board of Directors.

With his expertise in new in-depth network architectures, innovative methodology for understanding the influence of data on models, and automated methods for evaluating AI model resilience, Zico is invaluable to the company's technical directors.

Zico will also join the Council Safety and Security Committee with other directors including Bret Taylor, Adam D'Angelo, Paul Nakasone, Nicole Seligman, and Sam Altman (CEO) as well as OpenAI technical experts.

The committee is responsible for making recommendations regarding safety and security decisions that are important for all OpenAI projects.

"Zico adds a technical understanding and an in-depth perspective in terms of AI security and resilience that will help us ensure artificial intelligence is generally beneficial for all mankind," Bret Taylor, Chairman of the OpenAI Council, said in his official blog.

At First glance about Zico Kolter

Zico Kolter is a Computer Sciences Professor and head of the Department of Machine Learning at Carnegie Mellon University, where he has been an important figure for 12 years.

Zico completed his Ph.D. in computer science at Stanford University in 2010, followed by post-doctoral scholarships at MIT from 2010 to 2012.

Throughout his career, he has made a significant contribution to machine learning, writing many award-winning papers at prestigious conferences such as NeurIPS, ICML, and AISTATS.

Zico's research includes the development of the first method to create a deep learning model with guaranteed resilience. He also spearheaded techniques to embed hard constraints into AI models using classical optimization in neural network layers.

Recently, in 2023, his team developed an innovative method to automatically assess the safety of large language models (LLMs), which demonstrates the potential to bypass existing model protection through automated optimization techniques.