New OpenAI tool renews fears that “AI slop” will overwhelm scientific research

OpenAI has launched Prism, a groundbreaking AI-driven workspace tailored for scientists, which has quickly sparked skepticism within the research community. Many experts worry that this tool may exacerbate the ongoing deluge of subpar research flooding academic journals, a phenomenon they refer to as "AI slop." Prism is designed to assist researchers by providing a writing and formatting platform built on OpenAI’s GPT-5.2 model, integrated into a LaTeX-based text editor. This allows scientists to draft manuscripts, generate citations, and create diagrams from sketches while collaborating with co-authors in real time. The tool is available for free to anyone holding a ChatGPT account. Kevin Weil, vice president of OpenAI for Science, expressed his belief that 2026 will mark a pivotal moment for AI in scientific research, paralleling the transformative impact AI had on software engineering in 2025. He highlighted that ChatGPT handles approximately 8.4 million messages weekly concerning hard science topics, underscoring the growing role of AI in the scientific workflow. Prism is built on technology acquired from Crixet, a cloud-based LaTeX platform. OpenAI aims for this tool to alleviate researchers from the tedious aspects of formatting, allowing them to focus more on substantive scientific inquiry. In a demonstration, an OpenAI representative showcased how the software could seamlessly integrate relevant scientific literature and format bibliographies. However, the introduction of such advanced tools carries inherent risks. The ease of producing polished manuscripts through Prism could lead to an influx of submissions that do not significantly contribute to their respective fields. As the threshold for generating research-like content diminishes, the mechanisms for critically assessing such work remain unchanged, raising concerns about the integrity of the peer review process.