Two research teams independently solved the same open quantum cryptography problem using OpenAI's GPT-5.6 model, submitting their findings to arXiv.org just three hours apart. The teams, led by MIT PhD student Seyoon Ragavan and professors Prabhanjan Ananth (UC Santa Barbara) and Amit Sahai (UCLA), used the same AI model but approached the challenge differently. Both focused on 'unclonable encryption,' a cryptographic method that relies on quantum properties to secure data. Their work has sparked discussions about the role of AI in modern research and what constitutes an independent discovery in an era of shared tools.
The rapid pace of discovery highlights the growing influence of AI in scientific research. According to Prabhanjan Ananth, if a researcher encounters an open problem, the first step is often to check whether GPT-5.6 can solve it. Ragavan echoed this sentiment, stating that his current research methods have evolved significantly compared to those from two months ago. The use of AI is reshaping how mathematicians and cryptographers approach complex problems, with reactions ranging from enthusiasm for new possibilities to concerns about the erosion of traditional research identity.
The incident underscores the broader implications of AI in academia. As researchers gain access to powerful models like GPT-5.6, the line between independent discovery and collaborative problem-solving becomes increasingly blurred. The teams are now considering merging their papers, raising questions about the nature of originality in an AI-driven research landscape. Source: thedecoder
