AI Writes Most New Web Articles. Here's How Writers Win.
More than half of new articles on the web are now AI-generated, according to a recent analysis. That explains why so much content feels generic, fast, and strangely familiar. The question for working writers isn't "Is AI taking over?" It's "What work will still be worth paying for - and how do I own that space?"
Skip the extremes
Back in the 1960s, cultural theorist Umberto Eco described two camps reacting to new media: the apocalyptics (doom) and the integrated (hype). Both miss the point. The real story is how people use the tool, what incentives it creates, and who benefits.
We saw this with deepfakes during the 2024 election. Lots of warnings, real risks, and widespread overconfidence in spotting fakes. Post-election reviews suggest they intensified distrust and noise, but there's no proof they changed final outcomes. The sky didn't fall - it just got foggier. Learn to work in the fog.
What AI is actually writing
The bulk of AI-written pieces are low-stakes, formula-heavy formats: listicles, short news updates, how-tos, lifestyle blurbs, reviews, and product explainers. These exist to inform or persuade fast, not to break new ground. AI is built for this.
That's also where many freelancers used to get steady gigs: blog posts, SEO pages, social copy, basic translation. Those tasks now pay less, happen faster, or vanish. Treat that as a signal, not a sentence.
Collaboration is the default
Plenty of pros already co-write with AI: draft a few lines, expand, then rewrite. It saves time, but it can also pull your voice toward bland defaults if you aren't careful. That's the trade-off.
There's another wrinkle: models are increasingly trained on mixes of human writing, AI output, and hybrid text. Some researchers worry this loop can reduce originality and accuracy over time. Don't build your craft on the same generic phrasing the models feed themselves.
The real risk: sameness
Studies show AI can make writers feel more creative, yet it often narrows the range of ideas and smooths style into a familiar pattern. You've seen that tone - safe, polite, full of hedges and filler. It reads fine and lands flat.
There's also a cultural bias: models skew to Western, especially English-speaking, norms. Call it AI colonialism. If every piece sounds like the same "polished" blogger, original voices get rinsed out.
A practical playbook for writers
- Pick work with stakes. Lean into reporting, analysis, interviews, lived experience, and specialty knowledge. Commodity topics invite commodity rates.
- Build proof-of-work. Maintain beats, sources, field notes, datasets, and firsthand tests. Receipts win briefs and briefs win retainers.
- Guard your voice. Create a personal style guide: favored verbs, banned clichés, sentence rhythms, and examples of "this is me." Use it to edit AI output back into your tone.
- Use AI with constraints. Let it brainstorm angles, outline options, or produce raw clay. Then rewrite from scratch. Don't publish the first pass. Set prompts that include your rules and red lines.
- Show your work. Quote sources, link evidence, add screenshots, field photos, or test results. Transparency beats vibe.
- Specialize where accuracy matters. Health, finance, law, B2B tech, climate, policy. Clients in these areas care about risk, nuance, and accountability.
- Productize and price on outcomes. Monthly briefs, editorial calendars, conversion pages, content systems. Charge for strategy, not just word count.
- Create what machines can't. On-the-ground reporting, expert interviews, proprietary surveys, long-term series, transformative edits.
- Own distribution. Newsletter, community, or a focused LinkedIn/Twitter presence. Don't rely on platforms that favor volume over value.
- Upskill fast. Learn prompts, revision workflows, and tool stacks that actually save hours. Curated picks for copywriters are here: AI tools for copywriting and role-based options here: courses by job.
For editors: quick tells of machine-heavy prose
- Repeating sentence templates and transitions ("In conclusion," "Additionally," "On the other hand").
- High polish, low specificity: clean grammar, vague claims, no names, no numbers.
- Lists that feel comprehensive but miss obvious edge cases.
- Confident statements without linked sources or with circular citations.
- Hedging everywhere yet no concrete stance.
Detection tools are unreliable. Instead, ask for process notes, sources, and drafts. Great writers show their trail.
Deepfakes are a cautionary tale
People tend to overestimate their ability to spot synthetic media, and confidence often beats accuracy. That gap is why trust erodes. If your work is to persuade or inform, earn trust on the page: evidence, clarity, and a voice readers recognize.
If you want a quick primer on the tech side, see deepfakes. But the fix isn't technical - it's editorial rigor.
The upside
As generic content floods the web, original voice and real expertise become easier to spot - and to pay for. Expect slower headline growth from big models and higher demand for writers who bring sources, discernment, and a point of view.
AI will write more. Let it. Your edge is work with depth, receipts, and a style no template can mimic. Double down on that and you'll be fine - better than fine.
Your membership also unlocks:
