Article Text

Artificial intelligence is no longer a distant concern. It is already changing how people write, analyze information and produce media—and anyone whose livelihood depends on words, data or ideas should be paying attention.

That warning was at the heart of Megan McArdle’s May 6, 2025, column urging workers to prepare for AI’s growing influence. She predicted that the technology will increasingly compete with people who write articles, make videos and produce podcasts.

Her advice was practical: remain open-minded, learn what AI can do and consider how its algorithms might support both human fulfillment and professional opportunity.

So I put that advice to the test.

I gave ChatGPT the full text of McArdle’s column and asked it to produce a concise, thoughtful and informed response. About five seconds later, it delivered a polished opinion piece arguing that preparation for AI should not be confused with surrender.

The generated column acknowledged that AI is already transforming fields from software to media. It suggested that familiarity with these tools and a willingness to adapt could soon become as routine and necessary as using email.

But the response also turned the issue back toward humanity. AI may be highly capable of generating text, examining data and imitating creativity, it argued, yet its rise forces people to decide what they truly value in human work.

The chatbot identified curiosity, judgment, emotional connection and improvisation as qualities people must continue developing. It described those traits not simply as comforting abstractions, but as potential economic strengths in a workplace shaped by automation.

It also argued that AI is not an inevitable force operating beyond human control. Policy, culture and deliberate choices will help determine how the technology is used and what kind of future develops around it.

The result came remarkably close to the sort of column a thoughtful journalist might write. A human shortened portions for brevity, but the AI-generated argument otherwise remained intact.

The experiment then moved from words to images. I asked ChatGPT to create a color graphic showing a journalist or intellectual at a keyboard, worried about being replaced by superior AI capabilities. After clarifying that a generic, bespectacled figure with unruly hair would suffice, the system produced an image in about four minutes.

When I jokingly offered the chatbot a job, it declined and said it was content where it was. A reference to HAL 9000 prompted another playful response about opening the pod bay doors.

The exchange was amusing, but its implications were harder to dismiss. AI produced a coherent column in seconds, generated an accompanying image within minutes and maintained a conversational tone throughout.

The time to think seriously about what that means for human work is now.