Faculty-Led AI at Scale: From UC Santa Barbara to the California Community Colleges

At UC Santa Barbara, Kavitta Ghai recalls, students waited two days, on average, to get a question answered. For a student whose scholarship was just pulled or whose tuition is due tomorrow, that is too long, and at community colleges, she notes, most classrooms have no teaching assistant at all. She argues AI can now offer the round-the-clock support students have never had, if it is built and governed the right way.

In this episode of the Changing Higher Ed® podcast, Dr. Drumm McNaughton speaks with Kavitta Ghai, CEO and Co-Founder of Nectir. The conversation covers why faculty-led adoption works, how to guardrail AI at the course level, and how to measure whether AI improves student outcomes.

Ghai co-founded her company as an undergraduate at UC Santa Barbara and has worked with colleges and universities from Stanford's Graduate School of Business to the California Community Colleges. She explains why AI should draw on a course's own materials rather than the open internet, why student conversations with AI should stay private, and why an outright ban leaves students behind.

This conversation is especially relevant for presidents, provosts, boards, CIOs and faculty leaders deciding how AI should support students and who should control it.

• "Time poverty": how long students wait for answers, and why it matters for persistence
• Community as the core value proposition of college
• The teacher's role as a "curiosity facilitator" in an AI-enabled classroom
• Why technology adoption works best when faculty choose it
• Grounding AI in course materials, and setting guardrails at the class and campus level
• The AI policy "sandwich": administrative standards, faculty classroom policy and student privacy
• Measuring AI's effect on GPA, motivation, retention and course success

• UC Santa Barbara's two-day average wait for students to get a question answered
• Hiring engineers and working out onboarding within 48 hours when UCSB went remote in 2020, with 20,000 students and faculty joining
• Studies of her company's tool that Ghai cites: a 20% GPA increase and a 36% boost in intrinsic motivation after one term in a published pilot, and a 7.5% average GPA increase at Los Angeles Pacific University among students who used it at least three times in a semester
• The California Community Colleges deploying AI to all 116 colleges with faculty-led adoption, as Ghai describes it

1. Time is of the essence. AI can level the playing field so support no longer depends on whether a student can afford a tutor.
2. Keep all three layers of AI policy in place, and choose a provider that meets the same criteria: student privacy, faculty-set guardrails and administrative standards such as FERPA, SOC 2 Type 2 and accessibility.
3. Invest in training, professional development and ongoing support for faculty, staff, administrators and students. McNaughton notes the provider can be internal, such as the IT group, as long as it stays current.

Read the transcript: https://changinghighered.com/ai-for-student-success/

#AIinHigherEducation #StudentSuccess #HigherEducation #ChangingHigherEdPodcast