'From biology to black holes, ChatGPT is accelerating research': OpenAI really wants you to use ChatGPT as a research collaborator - and claims 8.4 million messages are sent every week on science and math (www.techradar.com)

OpenAI has revealed that its ChatGPT is increasingly being utilized as a research collaborator, with 8.4 million weekly messages focused on advanced science and mathematics exchanged by approximately 1.3 million users. This represents a nearly 50% growth over the past year, indicating a shift from experimental use to a regular component in research workflows. The integration of GPT-5.2 models has been particularly noteworthy; they can follow complex reasoning chains, verify their findings, and engage with formal proof systems, achieving commendable results in significant mathematical challenges, including contributions to open Erdős problems. The implications of this research-driven engagement are significant for the AI/ML community. By automating routine tasks such as coding, literature reviews, and experimental planning, researchers can expedite their workflows. GPT-5.2 has shown impressive accuracy rates in graduate-level benchmarks across disciplines, and its hybrid usage with specialized tools in chemistry and biology offers a robust framework for scientific exploration. However, while these advancements expedite research processes, OpenAI acknowledges the necessity for independent validation to ensure the durability and applicability of these results in making substantial scientific contributions.