How Santa Clara University’s AI Kitchen Builds Practical AI Skills Beyond Chatbots

The Problem With “AI as Chatbot” Thinking

When most people encounter AI tools for the first time, they encounter ChatGPT. That first impression tends to stick. The mental model becomes fixed: AI is a text box you type questions into.

Kai Lukoff, assistant professor of computer science and engineering at Santa Clara University, sees this as a genuine limitation — not just for individuals, but for institutions. “A lot of folks are sort of still stuck in this paradigm of AI as a chatbot,” he said. The newer generation of agentic AI tools, which can plan, execute multi-step tasks, and interact with external systems, represents a fundamentally different capability set. Most people never reach that layer.

AI Kitchen was designed to close that gap — not through coursework, but through structured, hands-on experimentation.

What AI Kitchen Actually Is

Every Friday, roughly 50 participants gather for nearly four-hour sessions at Santa Clara University. The group is deliberately mixed: computer science students, anthropology staff, Silicon Valley professionals, faculty from across departments. The sessions are intentionally “code-light,” meaning no programming background is required to participate meaningfully.

The format is workshop-style. Participants test emerging AI tools, discuss real-world applications across fields — medicine, marketing, education, real estate — and bring their own professional contexts into the room.

That cross-disciplinary friction is the point. When an anthropology department staff member joins a session alongside a CS major, the conversation shifts from technical implementation to societal implication. Both perspectives sharpen each other.

Why the Name “AI Kitchen” Is a Design Decision

Lukoff chose the kitchen metaphor deliberately, grounded in research on how physical and conceptual framing affects who feels welcome in technical spaces.

“If you create a space that is kind of stereotypical tech bro culture, a lot of women and underrepresented students will opt out,” he explained. The kitchen signals something communal, creative, and familiar — the opposite of an “AI hacker space.”

That framing had measurable effects. Tiffany Le, an incoming third-year computer science and engineering major, initially found the tech industry’s male-dominated culture alienating. AI Kitchen’s atmosphere changed her calculus. She eventually became the program’s student studio lead, and now actively recruits other students by leading with the same inclusive message that drew her in.

Inclusion here is not a stated value — it is an architectural choice embedded in the program’s name, format, and facilitation style.

From Workshops to Real-World Projects

The practical orientation of AI Kitchen has generated tangible outcomes beyond the sessions themselves.

One example: a staff member from the university’s donor relations office attended a session and identified a potential AI application — matching donors with faculty members based on shared interests and priorities. That idea became an active student project, developed collaboratively between AI Kitchen participants and the donor relations team.

This is the workflow pattern worth noting for anyone thinking about AI adoption in an organizational context:

AI Fluency vs. AI Literacy: A Meaningful Distinction

Lukoff draws a deliberate line between two concepts that are often conflated.

AI literacy means understanding what AI tools are and how they generally work. AI fluency means being able to pick up a tool, test it against a real problem, assess its strengths and limitations, and integrate it into actual work.

The distinction matters because literacy without fluency produces informed observers. Fluency produces practitioners. AI Kitchen is explicitly oriented toward the latter — which is why the sessions run for nearly four hours and involve hands-on exercises rather than lectures.

For founders, marketers, and operators thinking about AI adoption in their own organizations, this framing is directly applicable. Reading about AI tools is not the same as developing the muscle memory to use them effectively under real conditions.

Critical Engagement, Not Cheerleading

One of the more intellectually honest aspects of AI Kitchen is what it refuses to do.

Lukoff is explicit that the program is not designed to sell participants on AI. “I’m by no means a cheerleader for AI tools,” he said. Hesitancy among students, faculty, and staff is acknowledged rather than dismissed.

His argument for engagement is pragmatic rather than promotional: these tools exist, they are being developed, and universities are better positioned than most institutions to bring critical perspective to that development. Ceding that space to actors without that critical lens is itself a choice — and not a neutral one.

This framing is worth adopting more broadly. The most durable AI fluency is built on honest assessment of what tools can and cannot do, not on enthusiasm for the technology itself.

What Higher Education — and Organizations — Can Learn From This Model

AI Kitchen is a specific program at a specific university. But the design principles it embodies are transferable.

Remove the prerequisite barrier. Code-light sessions allow domain experts from any field to participate. The goal is applied judgment, not technical certification.

Mix the room deliberately. Interdisciplinary friction surfaces use cases and implications that homogeneous groups miss. An anthropologist and a software engineer will ask fundamentally different questions about the same tool.

Anchor to real problems. The donor relations project emerged because a real stakeholder with a real problem was in the room. That proximity between problem and experimentation accelerates useful learning.

Build critical capacity alongside practical capacity. Participants who understand tool limitations are more effective adopters than those who are simply enthusiastic.

The Takeaway

AI Kitchen is not a course. It is not a certification program. It is a structured environment for developing the kind of practical, critical AI fluency that most formal education programs do not yet deliver.

The soft vegetables thrown at the end of each session are a small signal of something larger: this is a space where people feel comfortable enough to play, to fail, and to learn without the weight of performance. That psychological safety is not incidental to the program’s effectiveness — it is central to it.

For anyone building AI capability inside an organization, the lesson is straightforward. The tools matter less than the environment in which people learn to use them.