Related ToolsChatgptClaude

Columbia's "AI-First" Writing Class Flips the Script on Campus AI Bans

AI news: Columbia's "AI-First" Writing Class Flips the Script on Campus AI Bans

Over half of U.S. teens already use AI for schoolwork, according to recent research. Most universities are still debating whether to allow it. Columbia University is running a class that requires it.

The course is called "Writing AI," and freshman Maximilian Milovidov - also a member of TikTok's Youth Council - wrote about his experience in it for NPR. The class flips the typical university approach: instead of policing ChatGPT usage, students bring their own ideas and outlines, feed drafts into chatbots, then document and critique what comes back. They accept or reject each suggestion with reasoning. The AI becomes a sparring partner, not a ghostwriter.

The most interesting takeaway is counterintuitive. Milovidov found that working closely with AI actually made him value his own writing more, not less. The chatbot output was bland. It "amplified our weaknesses back at us," he wrote, producing generic prose that made his "own messy and imperfect paragraphs" look better by comparison. His conclusion: "What distinguishes us is not whether we can produce text, but whether we can think, judge and revise."

The Practical Argument

Milovidov pushes back on framing AI in education as a binary choice. "Children born today will never know a world without AI," he wrote. The question is not whether students will use these tools - they already are, in large numbers - but whether they will learn to use them critically.

Research cited in the piece backs this up: moderate AI use during lectures improves student performance compared to both full automation (letting AI do everything) and minimal support (barely using it). The middle path works best.

What This Means for AI Tool Users

This mirrors what many professionals are discovering on the job. The people getting the most from tools like ChatGPT and Claude are not the ones who paste in a prompt and accept the first output. They are the ones who iterate, push back, and know when the AI is giving them filler.

Columbia's approach treats AI literacy as a skill worth teaching deliberately, not something students should figure out on their own or be punished for attempting. That is a more honest position than most institutions have taken so far.