GPTZero welcomes Dr. Dongwon Lee, commits to standard benchmarks in AI detection

GPTZero is excited to announce the addition of Dr. Dongwon Lee, Professor and Director of the Ph.D. Program at the College of Information Sciences and Technology (IST) at Penn State University, as a faculty member and advisor. Dr. Lee brings with him a wealth of expertise and a distinguished record of research in the field of generative AI.

Dr. Lee's decision to join GPTZero stems from a shared vision and a mutual commitment to tackling some of the most pressing challenges in the realm of generative AI. His deep understanding of computational and socio-technical solutions, aligns with GPTZero's mission to drive responsible AI adoption in education.

"The work GPTZero is doing in the realm of generative AI is truly groundbreaking, and I am excited to contribute to their research efforts. This partnership represents a natural synergy, as my team at Penn State has been exploring various research angles on similar issues for years," expressed Dr. Dongwon Lee. At Penn State University, Dr. Lee is the director of the PIKE Lab, one of the foremost research groups tackling AI detection within the A3 project and has been leading the NSF-sponsored SysFake project since 2017, focusing on developing innovative strategies to combat fake news through computational and socio-technical approaches. His expertise in this domain positions him as a valuable asset to GPTZero's mission to promote academic integrity and combat plagiarism effectively.

"Together, we are poised to make significant strides in advancing the field of generative AI while reaffirming our commitment to responsible innovation and academic excellence. This collaboration marks a significant milestone as we become the first AI detection company to partner with an academic lab," stated Alex Cui, Chief Technology Officer of GPTZero. 

While one of the strongest elements of GPTZero models is that we have trained our models for practical use cases and robust coverage across different writing domains. With the PIKE Lab, we discussed that there is a need for objective evaluations of AI models, as many AI detection providers only perform accurately on specific datasets and writing domains. One of GPTZero and PIKE Lab's first initiatives will be developing a dynamic benchmark for AI detection models — an objective benchmark not gameable by AI detection model providers. This will provide an objective evaluations of the success, coverage and scope of different AI detection models.

We're excited to develop an industry standard benchmark alongside the PIKE lab. Stay tuned for updates on what we accomplish together!