Artificial intelligence no longer lives only in engineering departments. It writes first drafts of legal briefs, triages medical scans, helps renovate old video games, and sometimes quietly convinces a customer service bot to give away a free refrigerator. As AI spreads into every profession, Duke University has made a bold argument: understanding these tools should not be a niche technical specialty. It should be part of the foundation of a modern education. That is the idea behind Duke AI for Everyone.
What is Duke AI for Everyone?
Duke AI for Everyone is a cross-disciplinary educational initiative at Duke University. Instead of focusing on advanced model building, it offers a structured path for students, faculty, and working professionals to become fluent in how AI systems behave, how they are built, and where they can fail.
The program looks less like a traditional computer science sequence and more like a liberal arts requirement. A nursing student might take a module on why predictive algorithms underperform for certain demographic groups. A political science major might train a simple classifier and watch bias sneak into the training data. Each course is hands-on, but none assume a background in coding. Duke calls this “democratising AI,” and the name is intentionally direct.
Who is it for?
Duke designed the curriculum for a wide audience. Current undergraduates from all nine schools are eligible, but the university also opened certain tracks to graduate students, staff, and not-for-credit online learners.
If you are a business analyst who wants to understand what your data team is actually doing, a high school teacher trying to build lesson plans around generative AI, or a retiree who simply wants to spot deepfakes, the non-technical path is relevant. Some courses sit behind Duke’s tuition wall, but several public lectures and the introductory module are available without enrolling at the university.
Inside the program’s core curriculum
Early pilot courses carry the AI for Everyone banner, and the schedule changes depending on the department. Still, most students begin with the same foundational class: AI Literacy and Human Values.
The introductory track
In the first two weeks, students learn the vocabulary that dominates AI headlines: large language models, training data, neural networks, and generative adversarial networks. There are no equations beyond basic probability. Instructors want students to understand what these systems can and cannot do, not to build one from scratch.
The final module asks students to draft a set of principles for responsible AI use inside their own future profession. Some of the best projects have come from unusual majors.
Domain-specific modules
Duke AI for Everyone does not stop at a generic survey. The medical school runs a module showing residents how AI triage tools hide unfairness. The law school explores copyright and liability questions raised by generative models. The business school tests AI agents that negotiate contracts, then breaks down what happens when the agents try to bluff.
By the end, students are expected to hit four specific milestones:
- Build and interrogate a real AI model without writing code.
- Explain the limitations of a machine learning system to a non-expert.
- Evaluate a real-world AI failure and identify its root cause.
- Develop a personal code of conduct for using AI in your own work.
What makes it stand out from online AI courses
There are plenty of “AI in 10 days” courses online. Some are genuinely helpful. Duke AI for Everyone distinguishes itself by creating a shared learning experience across the whole institution, not just a standalone certification.
Students consistently mention that the cohort-based seminars change their perspective. You are not alone in front of a screen; you are arguing with a history major about algorithmic bias while a statistics PhD student patiently corrects both of you. That kind of friction rarely happens in a typical online course.
A second difference is the emphasis on ecosystems. Machine learning models do not exist in a vacuum. They are developed, funded, and deployed by companies that sometimes stumble publicly. Consider what happened when password manager 1Password funded a Linux project and ended up wading into a right-wing mess. Everyone in that situation underestimated how quickly trust can evaporate. Duke’s program teaches students to think about the whole system around a piece of software, not just the algorithm inside it.
Why this matters beyond Duke
The initiative is creating a template for other universities. Most schools still treat AI as something that belongs in a computer science building. Duke has shown that a philosophy major can learn enough about model evaluation to ask the right questions in a boardroom, or a public health graduate can spot a dataset that will never represent the patients she serves.
That kind of literacy has become a workforce requirement. Employers report that they are not looking for every candidate to be a prompt engineer. They want people who know when to trust an AI output, when to challenge it, and how to explain those decisions to a sceptical boss. Duke AI for Everyone is built specifically around those skills.
How to get involved
If you are a Duke student, the short answer is to search the course catalogue for labels like AI 120, AI 130, and AI 140. The program also runs drop-in labs and a public speaker series each semester. If you are not at Duke, several of the foundational modules are posted as open-access materials, including lecture videos and project prompts.
For those who want a fuller credential, Duke AI for Everyone occasionally opens seats to external learners through the university’s online learning platform. The next cohort usually forms in early spring, and applications do not require a technical degree or a coding portfolio. You simply need a willingness to question what the machine tells you.
AI is becoming a shared skill this decade, much like reading and writing became shared skills in centuries past. Programs like this one are early experiments in how to teach that skill at scale. Even if the name changes or the courses evolve, the goal will stay the same: making sure the people who live with AI are the people who understand it.