The Convergence and the Effective Integration of AI Policies and Practices into California Community Colleges

From The Desk Of FACCC President Sarah Thompson

California’s Community Colleges have seen remarkable change over the last seven years. Across the system, we see radical shifts in how colleges generate revenue, shifts in the labor force, a bombardment of new technologies, and multiple state initiatives both guiding and constraining our system’s choices in responding to this “convergence”.

As a consequence, existential questions have arisen: What is a community college education? And how do we measure success in educating? Across the nation, we have seen grade inflation, rising academic dishonesty driven by large language model programs, and faculty wondering whether grades are even good indicators of learning.

How have we come to lose faith so quickly? In this article, I intend to raise awareness of the shock the system has taken and how any effort to transform our educational practices should be thoughtful about our academic, professional, and ethical priorities. Rushing to integrate artificial intelligence (AI) into every class and department can create more problems than it solves. We need breathing room to evaluate, restructure, and perhaps even roll back some of our AI initiatives. This is going to take more than attending a workshop.

Since 2019, we have experienced what I call the “convergence”, the intersection of major structural, technological, and mission changes within our system. These include, but are not limited to, the growing dependence on Distance Education (DE) for FTES generation; a greater reliance on part-time faculty; the 2030 vision for dual enrollment and credit and noncredit workforce development, which could further increase reliance on both DE and part-time faculty; the AI industry’s growing presence and expansion into higher education; mission-changing initiatives such as AB 1705 (Irwin); and the emergence of California Community College (CCC) baccalaureate degrees. Any successful AI transformation within our system must account for these factors. We will need to address our system’s weaknesses, capitalize on our strengths, and create short- and long-term solutions.

The Great Distance Education Migration

Throughout the 2000s and 2010s, we saw a steady but incremental rise in asynchronous distance-learning classes. In 2019, every college in the system had some type of DE offerings, and Calbright, the system’s all-DE college, was slowly finding its footing. In this context, DE was seen as an alternative for students who did not fit the “traditional” face-to-face student model. The system viewed it as an attempt to be inclusive and welcoming to students facing other challenges in their time, work, childcare, eldercare, etc. Colleges across the state offered DE, but they depended only moderately on the FTES these courses generated. Eighty-one percent of colleges in the system generated less than 25% of their total FTES through distance education.

The pandemic sparked a shift in most districts’ dependence on DE enrollments. Like the annual wildebeest migration, the CCCs charged toward an all-DE platform, imperfectly implemented but assumed temporary. However, as of Fall 2025, most colleges in the state have not come close to their pre-pandemic in-person enrollments. In fact, only 8% of colleges in the system now have low or moderate reliance on DE enrollments, while 34% of colleges are highly or extremely reliant on DE enrollments.

Seven-year comparison – college dependence on distance education enrollments

Term	Under 10% Reliance “Low”	10-25% Reliance “Moderate”	25-50% Reliance “Average”	50-75% Reliance “High”	75% + Reliance “Extreme”
% of Colleges Reliant on DE Enrollments as Part of Their Total Fall 2019	N=8 7%	N=85 74%	N=18 16%	N=2 2%	N=2 2%
% of Colleges Reliant on DE Enrollments as Part of Their Total Fall 2025	N=2 2%	N=7 6%	N=68 58%	N=37 31%	N=4 3%

The struggle to reinvigorate on-campus culture through in-person enrollments is real. And costly. Unlike Calbright College, which was created as an online-only education center, all the other colleges have physical spaces and personnel designed to support a functional face-to-face educational experience. Even if 70% of their enrollments are online, colleges must still maintain their physical campuses and build an effective online support system.

The chart below shows how much each campus changed. Regardless of a college’s 2019 level of dependency, the question is how much of an enrollment shift it experienced by 2025. If we consider student enrollments as cost centers, the more extreme the change in dependency, the more difficult it is to balance the costs of the physical infrastructure and face-to-face employees with the growing needs of online enrollments – providing student support, round-the-clock tutoring, filtering out fraudulent enrollments, tools for securing the academic honesty of its enrollees, etc. It also means that any attempt to recreate a mostly in-person campus and educational environment could be viewed as financially risky, either because of fear of losing online students to other campuses or because the online dependency includes students from outside the district, so a face-to-face transition would mean losing FTES due to proximity.

Impact level	Low (under 10%)	Moderate (10–20%)	Average (20–30%)	High (30–40%)	Extreme (more than 40%)
% Overall and list of colleges	6%	20%	36%	24%	14%

(Four colleges reported less dependence on DE enrollments in 2025 than in 2019 and are not included).

See which colleges fall in each category

• Low: Allan Hancock, Calbright, Foothill, LPC, Miramar, Siskiyous, Cerro Coso
• Moderate: Bakersfield, Cabrillo, Chabot, Coalinga, Coastline, Crafton Hills, Cypress, Irvine, LA Trade-Tech, Lassen, Marin, Merritt, Monterey Peninsula, Ohlone, Orange Coast, Redwoods, Reedley, Mesa, Sequoias, Shasta, Taft, West LA
• Average: Alameda, American River, Barstow, Butte, Canyons, Citrus, Columbia, Compton, Contra Costa, Cuesta, Cuyamaca, Diablo Valley, Fresno, Fullerton, Glendale, Golden West, Grossmont, Hartnell, Laney, Lemoore, Long Beach, Los Medanos, Mission, Modesto Jr., Palomar, Pasadena, Rio Hondo, Riverside, Saddleback, San Bernardino, San Diego City, San Joaquin Delta, San Jose City, Santa Ana, Santa Barbara, Santa Monica, Santa Rosa, Santiago Canyon, Solano, Southwestern
• High: Antelope Valley, Berkeley City, Cerritos, Copper Mountain, De Anza, Desert, East LA, El Camino, Evergreen, Gavilan, LA Harbor, LA Pierce, LA Valley, Lake Tahoe, Mendocino, Merced, MiraCosta, Moreno Valley, Mt. SAC, Napa, Oxnard, Porterville, Sacramento City, Sierra, Victor Valley, West Valley, Yuba
• Extreme: Chaffey, Clovis, Cosumnes River, Feather River, Folsom Lake, Imperial, LA City, LA Mission, LA Southwest, Moorpark, Norco, San Diego College of Continuing Education, City College of San Francisco, Ventura, Woodland

A mostly online student population brings challenges; it leaves the college more vulnerable than colleges with a larger on-campus presence to fraudulent enrollments, AI-based academic dishonesty, privacy violations, and other security issues such as the Canvas ransomware attack. Any potentially negative outcome of distance education will be amplified in districts with higher reliance on DE. It also becomes difficult to create the welcoming on-campus environment that so often promotes retention. Instead, facilities and sparsely populated spaces give a “ghost town” effect. We are more vulnerable as a system since we have the greatest fiscal dependency on DE in higher education. While our sister systems use the DE modality, their fiscal reliance is lower: a little under 1/3 of California State University (CSU) course offerings are online, and 1-6% of University of California (UC) course offerings are online.

The Ubering of Instruction

Similar to the vulnerability of our fiscal dependency to online modalities, the CCC system is also stymied by our overdependence on part-time instructional labor. We’ve seen an explosion of part-time and short-contract labor in the classroom over the past 40 years across the country, at all levels of higher education. This “ubering” of instruction, or having instruction delivered by largely exploited gig workers, introduces dysfunction into operations, despite the cost savings for colleges and universities. We face this challenge in California across all public systems, but, as with the reliance on DE, this structure is most pronounced in community colleges. Part-time and short-contract instructors comprise just under 50% of UC faculty members. Just under 50% of CSU faculty are part-time and short-contract. In the CCCs, part-time workers are the overwhelming majority of instructors, making up 69% of CCC faculty. And unlike the CSUs and UCs, the Community Colleges offer no opportunities for part-time faculty to attain full-time, non-tenure-track contracts.

The dysfunction of this dependency has been well documented in terms of student impact:

• less likely to hold office hours and more likely to teach what the system would consider “overload” since the compensation is insufficient;
• less likely to access or even be offered professional development;
• more likely to be disconnected from the college community, especially with the growth of distance education;
• likely to have less representation in campus governance or collective bargaining;
• less likely to feel that their academic freedom protections are assured; and
• more likely to be punished for negative student evaluations or student grievances.

When campuses need to make system changes that require shifts in instructional content, pedagogy, assessment, etc., the more dependent a campus is on part-time faculty, the harder it will be to ensure that information, training, and collaboration occur equally and effectively. Part-time faculty are harder to reach, schedule, and train because many work across multiple campuses or have exclusively online assignments. Professional development for part-time faculty is just as expensive as that for full-time faculty – there is no economy of scale. For example, a campus with only 20% full-time faculty will cost five times as much to train all faculty as it would to train only full-time employees. It may actually be more expensive, since most collective bargaining units offer hourly pay to part-timers who participate in out-of-instruction activities.

“This is going to take more than sending faculty to a workshop.”

FACCC President Sarah Thompson

The table below looks at the frequencies of part-time dependency across the system (CCCCO DATAMART):

Percent of part-time faculty out of total faculty	Percent of colleges falling in this category	Notes
Extreme dependency >= 80%	9%	The highest is 88%
Very high dependency 70-79%	33%
High dependency 56-69%	53%
The lowest range of dependency 47-55%	5%	Only two campuses have under 50%

This dependency has varied statewide since 2019 (pre-COVID):

Pattern of FT/PT employees by college	Percent of colleges falling in this category	Notes
Colleges saw gains in both FT and PT faculty numbers	19%	One college is excluded as it was not an independent college in 2019
Colleges saw gains in FT, losses in PT faculty numbers	31%
Colleges saw gains in PT, losses in FT faculty numbers	19%
Colleges saw losses in both FT and PT numbers	31%

Part-time faculty are not equally distributed around the system. Some academic sectors have significant overrepresentation of part-time employees: non-credit/workforce development and dual enrollment. The challenge of this FT/PT inequity is currently due to an ambitious effort to rapidly expand these programs across the state. As this occurs, the part-time population should increase even further. Any effort to bring professional development to these faculty populations will compound the difficulties outlined above.

The AI Blitzkrieg Over the Instructional Maginot Line

The need to respond to, adapt to, and, when appropriate, adopt the barrage of LLM/AI technologies is urgent. The integrity of academic pedagogy, assessments, and student learning has all been threatened through the rapid changes brought about by AI. But even though this turning point in higher education needs our attention and action, our decisions must be made with the cornerstone of higher education, academic freedom, protected and intact. While industry players are likely to immediately label any hesitancy to adopt AI as a “luddite” response, our governing decisions should be made based on what is best for the community we are serving and the integrity of our courses and, ultimately, our profession.

Several educational TK-12 sectors in “breaking news” have gone public with cautionary AI policies. Norway announced this summer that it is limiting or banning the use of AI in schools altogether. This adds to its current ban on cell phones in schools, which significantly limits tablet use and increases reliance on books. Their data reveals that AI usage is interfering with learning thresholds, making students more reliant on AI and less able to critically evaluate it.

American neuroscientist Jared Cooney Horvath’s research reveals similar outcomes. In his book Digital Delusion, he finds that high computer/tablet use correlates with poorer academic performance. This supports decades of research showing that students remember more from lectures when they take handwritten notes, and that they remember more from reading than from passively watching videos. Successful learning is a multisensory experience, and denying young learners the opportunity to develop these skills hinders them in the long run.

We are seeing these concerns about AI reliance interfering with skills acquisition at the other end of the educational spectrum, too. This summer, three prestigious law schools have sought to set their training apart by banning the use of AI by students and faculty in their programs: UC Berkeley, the University of Texas, and the University of Chicago. Chicago took it a step further by also forbidding laptops, tablets, and phones in classrooms for first-year students.

The onslaught of academic dishonesty documented by undergraduate professors has many advocating for a return to analog as well. The viral Inside Higher Ed article about the discoveries of Brown Professor of Economics Robert Serrano reveals the suspicions of many online instructors that their assessments are being corrupted by AI cheating. In Spring 2026, Dr. Serrano suspected that his students’ midterm grades were not authentic, as the class average was a whopping 96, with 40 students receiving a perfect score (over 20 points higher than previous classes). He then decided to hold an in-person cumulative final. Of 86 students, 18 immediately dropped the class, and nine did not show up for the final. The class average of those who did take the exam was a 49 — only a handful replicated the score from their midterm. Despite lowering the passing threshold, 19 students still failed the course. At Las Positas College in Livermore, Statistics students taking online instruction with online examinations demonstrate greater success and experience less attrition than students taking online instruction with in-person testing.

This is one example of a learning outcome that led the CSU Math Consortium to pass a resolution urging universities not to articulate CCC math courses without in-person testing. Current research by FACCC Think Tank Coordinator Michael Peterson asked CCC faculty: How confident are you in your students’ ability to succeed even if they technically passed?

If faculty themselves lack faith in their assessments as an indicator of student learning, it is not surprising that others outside the system do as well.

Even more concerning, AI seems to interfere with student learning even when it is used to support, rather than complete, coursework. Studies such as Bastani et al. (2025) show that less is committed to memory, leading to less specialized knowledge and an inability to recall information, especially in skill-building tasks. AI can create great efficiencies, but it is most helpful once expertise is already established, not while it is being developed.

This leads us to the essence of the future of higher education: balancing the integration of practical, innovative AI use with the cognitive and intellectual skill-building that will help students use AI effectively rather than be used by AI. This future will require all faculty to be AI literate, even if, for their discipline or skill-building task, it’s best to integrate more analog components into their pedagogy. The entry-level work sphere is shrinking due to the efficiencies AI provides. Somehow, we still need to produce competitive college graduates who can sidestep the traditional entry-level training environments and succeed at the higher levels historically reserved for those with more experience. This means our students need to be more creative, more adept at critical thinking, politically and socially savvier than previous generations, with a deep understanding of the governance and dangers of AI, and be effective users of the technology.

This is going to take more than sending faculty to a workshop.

References

1. EdSource. (2025). At community colleges, online classes remain popular years after pandemic.
2. University of California. Course Enrollment by Instructional Modality.
3. University of California. (2024). Accountability Report, Chapter 5.
4. California State University. Employees: Facts About the CSU.
5. EdSource. (2023). California community colleges rely too much on part-time faculty and misspend funds, audit finds.
6. Reuters. (2026). Norway imposes near-ban on AI in elementary schools.
7. Institute of Education Sciences. AI in K–12 Education: The Good, the Bad, and the Guardrails to Consider.
8. NPR. (2026). Article on AI in Schools and Education.
9. YouTube. Video reference on technology and learning.
10. Horvath, Jared Cooney. (2025). The Digital Delusion: How Classroom Technology Harms Our Kids’ Learning—and How to Help Them Thrive Again. LME Global.
11. Edutopia. Why Writing by Hand Beats Typing in 6 Charts.
12. Bastakiss. Reading vs. Watching Videos: A Comprehensive Analysis of Learning Efficiency.
13. Inside Higher Ed. (2026). AI-Proof Lawyers? Some Law Schools Restrict AI Use.
14. Bastani et al. (2025). Research published in Proceedings of the National Academy of Sciences on AI and student learning.