This blog, which provides an overview of AI Commons, is the first in a series focusing on the California State University’s (CSU’s) systemwide approach to generative AI (GenAI). Before writing this, I experimented with using ChatGPT to generate content. At the end of the blog, I’ll share with you how that experiment went and how much of the AI-generated “content” I used.
Overview
The CSU system is taking a proactive, systemwide approach to supporting the adoption of generative artificial intelligence (GenAI) on its campuses. In February 2025, the CSU Chancellor’s Office announced a robust initiative “to create an AI-empowered higher education system that could surpass any existing model in both scale and impact.” The CSU committed $16.9 million to the effort, including a new partnership with major technology companies. According to the CSU, no other university system is implementing GenAI at this scale.
In launching the initiative, CSU Chancellor Mildred García said that by adopting GenAI the CSU seeks “to elevate our students’ educational experience across all fields of study, empower our faculty’s teaching and research, and help provide the highly educated workforce that will drive California’s future AI-driven economy.” García also emphasized that access to and use of GenAI tools in the CSU will need to be “responsible and equitable.” To that end, the CSU made ChatGPT available to all its students, faculty, and staff across all campuses. The CSU also launched a digital hub, called AI Commons, to provide guidance, training, and resources on the use of GenAI technologies. As well as partnering with tech giants, the CSU is also partnering with its own faculty and campuses to support the development of AI-informed instruction, administrative functions, and workforce training.
Meanwhile, university leaders across the U.S. have expressed concerns about the implementation of GenAI in higher education, including the propensity for plagiarism in student work and academic papers, over-reliance on generative tools instead of human analysis and instruction, and the prevalence of false and hallucinatory information in GenAI results. Related concerns include data privacy and governance, the use of biased and unpredictable algorithms, vendor influence, and increased workload for faculty and staff in adapting to the use of GenAI in instruction and in broader academic and administrative roles. Beyond the education sphere, other social and justice concerns include the substantial environmental and other costs associated with energy and water use for data centers, pollution and higher costs of electricity (especially for low-income neighborhoods near the data centers), potential loss of jobs and livelihoods due to the widespread adoption of GenAI, and potential for data exploitation.
By supporting and adopting GenAI systemwide, the CSU is seeking to prepare its students for a GenAI-informed world by balancing innovation with responsible policy, GenAI literacy with academic integrity, and the use of GenAI tools with human analysis and leadership. The extent to which these aims can be achieved remains to be seen. To better understand the CSU’s approach, this blog describes the resources available on CSU’s AI Commons for students, faculty, and staff. Future blogs will examine the perspectives of some CSU students, faculty, and staff about GenAI.
GenAI Resources for Students
For students, AI Commons offers access to ChatGPT Edu, a version of OpenAI’s model that is customized for academic use, with added privacy and content controls. Beyond access to this tool, the portal at this stage in its development delivers rudimentary information to students. A policy document called “Ethical and Responsible Use of AI for Students” offers general principles of AI Use in academic settings, provides some examples of ethical and unethical uses, and directs students to engage with their instructors about GenAI generally and how GenAI tools should be used in their courses.
In addition, AI Commons offers several introductory asynchronous trainings in basic AI literacy. Most of these trainings were created by tech vendors, but the most substantial one, called “Academic Applications of AI,” was developed by San Diego State University (SDSU). The SDSU course offers a microcredential for students (with a commitment of 2 to 10 hours online). The course covers: how GenAI tools actually work; ethical and responsible use; appropriate and inappropriate uses in academic and career pursuits; and strategies for effective querying. According to SDSU, over 9,000 CSU students have logged onto the course, about a third of which have received a microcredential (badge).
GenAI Resources for Faculty
The resources provided for faculty on AI Commons, compared with those provided for students, are much more robust, which is reasonable, considering that the CSU is encouraging its faculty to instruct students about effective and responsible GenAI use within their discipline. As with the student portal, AI Commons provides faculty with direct access to ChatGPT Edu. In addition, most faculty have access to Google Workspace and/or Microsoft Office 365, both of which have GenAI features.
Based on guidelines from the systemwide Academic Senate (ASCSU), the CSU encourages all faculty to establish clear policies on whether GenAI tools can or cannot be used by students for coursework, and if so, under what circumstances and conditions. Faculty are also expected to explore with students the range of GenAI tools available and the implications of their use or non-use in relation to their discipline. To assist in these areas, AI Commons offers policy guidelines in the following areas:
- Guidelines for faculty regarding AI in instruction. These guidelines, developed and approved by ASCSU, are organized in three areas: for all faculty; for faculty who do allow the Use of GenAI; and for faculty who do not allow the use of GenAI.
- Ethical and responsible use of AI. These resources include “ETHICAL Principles AI Framework for Higher Education,” which were developed at CSU Fullerton and include a short bibliography. A second resource, “AI Literacy Literature Summary,” was developed at the University of Notre Dame and provides a summary of the most highly cited literature regarding AI literacy, including rubrics for assessing and evaluating AI literacy.
- ASCSU Resolutions on GenAI provides resolutions adopted since 2019 by the Academic Senate about GenAI.
As with the policy frameworks described above, the online professional development courses offered on AI Commons for faculty appear to be more robust than those offered to students. They include an Introduction to AI for Teaching and Learning, created by CSUMB; the Academic Applications of AI (AAAI) for Faculty 2.0, created by SDSU; two AI webinar series for faculty; and courses and books for ChatGPT available on CSU Learn.
GenAI Resources for Staff
The resources dedicated for staff on AI Commons are not as extensive as those for faculty. AI Commons provides all staff with direct access to ChatGPT Edu. Many staff, depending on their positions, also have access to GenAI features associated with Google Workspace and/or Microsoft Office 365.
AI Common’s policy guidelines for staff provide rules and restrictions about GenAI use related to: confidential data; information security protocols; content verification; phishing alerts; GenAI tool procurement; and user privacy. As with the faculty portal, AI Commons also offers staff “ETHICAL Principles AI Framework for Higher Education,” developed at CSU Fullerton.
The online professional development courses offered for staff include AI Tools for Higher Ed Professionals and Administrators, developed at Fresno State University. The course offers a credential and provides skills in evaluating GenAI tools for professional tasks and identifying specific areas where GenAI can be applied to streamline workflows or address challenges. Other courses include: A webinar for staff and access to the webinar series developed for faculty; AI trainings developed by Microsoft; and courses and books for ChatGPT available on CSU Learn.
Looking Ahead
It’s important to keep in mind that AI Commons is in its first year and that its resources will continue to evolve as, hopefully, the CSU system continues its support for faculty and campus projects to develop GenAI guidelines and professional development resources. In particular, SDSU has created a student survey instrument regarding student perspectives about AI that has been used by several CSU campuses. Our next blog will describe some findings from these survey instruments.
My Experiment Using ChatGPT for this Blog
For writing, I’ve used ChatGPT previously in limited ways to suggest alternative word choices, to shorten a passage, or to phrase a sentence, a paragraph, or an idea differently. I’ve also used it to help in drafting letters and emails. The results have been mixed, with minor errors and real clunkers stylistically, but also with some good suggestions that I’ve drawn from to broaden my language usage for issues I’ve been writing about for a long time.
For this project, I input a detailed prompt into ChatGPT asking for a 1,000 word blog (directed to an academic audience) about CSU’s strategic approach to GenAI and its resources on AI Commons for students, faculty, and staff. Included in the prompt were url addresses for AI Commons and for its student, faculty, and staff portals. Note that this prompt already provided a structure for the blog: a strategic approach followed by three sections.
In its response, ChatGPT drew primarily from news releases, was excessively laudatory of CSU’s “strategic pillars,” and was enthusiastically supportive of GenAI adoption. I asked ChatGPT to revise the blog to make it more neutral and to include challenges posed by AI use in higher education. I was impressed by the balanced approach of the second version, but it was dominated by lists of bulleted information that I later found to be simplistic. The second version also removed all the links that were in the first version. The process underscored what I’ve come to know about writing: I needed to read the material myself in order to understand what I thought about it and how I should describe it. I did not use either ChatGPT version. As a final note, however, both ChatGPT versions came in at less than 1,000 words, whereas this blog (at 1,275 words) is well over that goal.