After revolutionizing writing, coding, and conversational assistance, OpenAI is now turning its attention to another strategic pillar: scientific research. With the launch of Prism, the company is offering a free AI workspace designed for researchers, educators, and students, and directly integrated into the ChatGPT ecosystem. The goal is clear: to make AI a native tool for academic work, capable of automating the most time-consuming tasks while leaving intellectual control in human hands.
Prism, a new AI environment for research
OpenAI describes Prism as the scientific equivalent of what Claude Code is for developers. The tool is based on the GPT-5.2 Thinking model, designed to reason through complex problem chains, and on Crixet technology, a cloud platform specializing in LaTeX composition recently acquired by OpenAI. This combination allows Prism to function not only as a writing assistant but as a comprehensive work environment dedicated to academic writing.
Unlike many specialized tools, Prism is now free for all ChatGPT personal account holders, with unlimited access to projects and collaborators. The Crixet brand is thus being phased out in favor of native integration into the OpenAI platform.
LaTeX in natural language: a massive time-saver
In STEM fields and more broadly in academia, LaTeX is the standard for scientific writing. Its power comes with a steep learning curve, particularly when it comes to creating graphics, diagrams, or advanced equations. Prism tackles this challenge head-on.
Thanks to GPT-5.2, users can now describe in natural language what they want to produce—such as a neural network architecture diagram—and let the AI automatically generate the corresponding LaTeX code, including through complex libraries like TikZ. The diagram is then rendered visually without any manual intervention. According to OpenAI, this type of automation can reduce the time spent on the technical formatting of scientific articles by more than 50% 1.
A literature review and teaching assistant
Prism does more than just help with document composition. The tool acts as a research assistant, capable of scanning scientific literature related to the topic at hand, extracting relevant references, and automatically generating a formatted bibliography. This feature addresses a critical need, given that the volume of scientific publications roughly doubles every 12 to 15 years2.
The tool is also designed for the higher education sector. Prism can help structure advanced course syllabi, provide educational content, and generate exercises for complex subjects, including general relativity. This versatility positions Prism at the intersection of research, pedagogy, and academic publishing.
Automation with human oversight; responsibility remains with people
OpenAI emphasizes, however, that Prism does not absolve researchers of their responsibility. The risk of misattributions or contextual errors—often referred to as AI hallucinations—remains. Kevin Weil, OpenAI’s vice president of science, points out that the goal is not to delegate scientific validation, but to streamline workflows while maintaining strict human oversight.
This approach is in line with emerging best practices in AI research, where automation is viewed as a driver of productivity rather than a source of scientific authority3.
A deliberate democratization of scientific AI
By making Prism free for personal ChatGPT accounts, OpenAI is making a bold strategic move. Unlimited access to projects and collaborators significantly lowers the barrier to entry for students, graduate students, and independent researchers. In the longer term, OpenAI plans to roll out Prism to its Business, Team, Enterprise, and Education plans in order to integrate it into institutional environments.
This strategy is part of a broader trend. The use of AI tools in research is booming, with the market projected to exceed $9 billion by 20304. OpenAI clearly intends to play a central role in this.
Ethical Issues and the Transformation of Research
The arrival of Prism, however, raises fundamental questions. By automating formatting, bibliographic research, and certain aspects of the teaching process, AI risks standardizing academic output and reinforcing dependence on proprietary tools. The issues of source traceability, model transparency, and equitable access become crucial, particularly in countries and institutions with limited resources.
More broadly, Prism illustrates a profound transformation in scientific work. The challenge is no longer simply to generate knowledge, but to know how to intelligently leverage AI tools to explore more quickly, compare more broadly, and write more effectively. The research of tomorrow might thus be defined less by technical mastery of these tools than by the ability to maintain intellectual control over them.
Toward an Integrated AI Lab
With Prism, OpenAI has reached a symbolic milestone. AI is no longer just a peripheral assistant, but a central workspace integrated into researchers’ daily routines. If the promises of time savings and accessibility are fulfilled, Prism could become the de facto standard for a new generation of scientists—provided that rigor, verification, and critical thinking remain at the heart of the process.
Learn more
This transformation of search into a smart workspace is part of a broader evolution in tools for exploring and organizing information. On a related topic, check out our article “Google Search’s AI Mode Expands to 180 Countries: A New Era of Search”, which analyzes how search engines are gradually integrating generative AI features to move from simple queries to continuous cognitive assistance.
References
1. OpenAI. (2026). Prism and AI-assisted scientific writing.
https://openai.com
2. Bornmann, L. & Mutz, R. (2015). Growth rates of modern science. Journal of the Association for Information Science and Technology
https://doi.org/10.1002/asi.23329
3. Nature Editorial. (2023). AI tools in research integrity.
https://www.nature.com
4. MarketsandMarkets. (2024). AI in Research Market Forecast.
https://www.marketsandmarkets.com
