GPT-5.6 Helped Two Teams Solve the Same Quantum Crypto Problem Hours Apart

Two research teams independently solved the same open quantum cryptography problem using OpenAI’s GPT-5.6 Sol Ultra. According to the article, their papers were submitted to arXiv just three hours apart. The overlap makes the case a clear example of how powerful AI systems are becoming part of the research workflow, not just a tool used after the main insight.

The problem involved “unclonable encryption,” a method that relies on quantum properties. The article identifies MIT PhD student Seyoon Ragavan on one side and professors Prabhanjan Ananth of UC Santa Barbara and Amit Sahai of UCLA on the other. Both efforts used the same AI model but took different approaches, and the researchers are now considering merging their papers.

The case raises a practical question for research communities: what counts as independent discovery when multiple groups can ask the same AI model for help? Ananth is quoted as saying, “If someone mentions an open problem, the first thing is to see if GPT solves it.” Ragavan adds that the way he does research now has “nothing to do” with how he did research two months ago, underscoring how quickly AI-assisted methods are changing expectations.

The article frames the episode as part of a broader shift in math and science, where AI is increasingly influencing how open problems are attacked. Reactions range from excitement about new possibilities to concern over identity and credit in fields built around human originality. For readers tracking AI research, the key issue is no longer whether advanced models can assist discovery, but how institutions will evaluate authorship, priority, and contribution when they do.