Most "human PhD writers" are now just AI whisperers. How to tell if you're paying for ChatGPT.

[SamWilliams]

The industry has flipped. In 2022, you were worried about getting an ESL writer. In 2026, you're worried about getting a guy with 3 browser tabs open.

How they operate now:

1. Service charges you $50/page.
2. "Writer" pastes your prompt into Claude 3.7 or GPT-5.
3. Runs it through a "humanizer" tool (these are a real thing now).
4. Delivers in 4 hours.
5. Profit.

How to spot AI-generated content (even "humanized"):

• Over-explaining: AI hates ambiguity. It will explain the obvious in three different ways.
• Formulaic transitions: "Moreover, Furthermore, In addition, Consequently." Humans use "But," "Yet," "So."
• The "As an AI" slip: I've literally seen this left in a conclusion. They didn't even delete the system prompt.

The irony: You're paying a human to use AI so you don't have to. The cognitive dissonance is staggering.

 

[Willow]

Let me add some practical detection tips:

The "Too Perfect" Test: AI-generated text often has zero errors. No typos, no slightly awkward phrasing, no comma splices. Human writing has small imperfections. If a paper is grammatically flawless but intellectually hollow, be suspicious.

The Source Check: AI fabricates sources. It will create convincing-looking citations to non-existent journals. Always verify a random sample of the bibliography. If 20% of the sources don't exist, the whole paper is suspect.

The Voice Consistency Test: Does every paragraph sound like it was written by the same person? AI often shifts tone mid-paper.