Before and After: Humanizing a ChatGPT Essay

Here is a paragraph generated by ChatGPT in response to the prompt: "Write a paragraph for an essay on climate policy, discussing the economic arguments for carbon pricing."

Original: "Climate change is one of the most pressing issues of our time, and addressing it requires comprehensive policy solutions. Carbon pricing is widely regarded as an effective economic tool for reducing greenhouse gas emissions. By putting a price on carbon, governments can incentivize businesses and individuals to reduce their carbon footprint. Furthermore, carbon pricing generates revenue that can be used to fund renewable energy initiatives and support affected communities. Additionally, it creates a level playing field for businesses by ensuring that the environmental costs of production are reflected in market prices. In conclusion, carbon pricing represents a balanced approach to addressing climate change while supporting economic growth."

Word count: 113 words. AI detection score on multiple tools: 90–95% AI probability. Reading quality: generic, predictable, formulaic.

Before looking at the humanized version, it is worth identifying the specific problems — because these are the patterns you will encounter in your own AI drafts.

• Opens with a generic statement of importance ("one of the most pressing issues of our time") — this opening pattern is extremely common in AI academic writing
• "Furthermore" and "Additionally" as connective tissue — carrying no logical information about why the next point follows
• Every sentence is in the 15–22 word range — zero variation in rhythm
• No specific data: "widely regarded," "can incentivize," "generates revenue" — all hedged in ways that avoid any concrete claim
• Concludes with "represents a balanced approach" — the generic positive-framing AI conclusion

After: "Carbon pricing is the most economically coherent response to climate change — and also the most politically difficult. The logic is simple: when companies can emit carbon for free, they do. A carbon price changes that calculation, shifting costs from taxpayers to emitters. In the EU, where the Emissions Trading System has been running since 2005, covered sectors have reduced emissions by roughly 35% while the scheme generated €38 billion in member-state revenue in 2022 alone. That revenue matters: it is what funds transition support for coal workers in Poland and energy efficiency subsidies for households in France. The hard part is not the economics. It is the politics — which is why carbon pricing has failed in more legislatures than it has passed."

Word count: 128 words. AI detection score: 12–20% AI probability across the same tools. Reading quality: specific, opinionated, engaging.

Change 1: Open with the argument, not the context

"Carbon pricing is the most economically coherent response to climate change" replaces the generic "Climate change is one of the most pressing issues." The humanized version opens with a claim — specifically a strong one — rather than a scene-setting statement. Human writers argue; AI writers contextualize.

Change 2: Add the tension immediately

"and also the most politically difficult" adds the tension that the original paragraph ignores entirely. The original presents carbon pricing as straightforwardly beneficial. A human writer covering this topic knows it is contested, and the strongest essays acknowledge that upfront.

Change 3: Explain the mechanism simply

"when companies can emit carbon for free, they do" replaces the abstract "incentivize businesses to reduce their carbon footprint." The short, direct explanation is more convincing than the policy language.

Change 4: Add real data

The EU ETS example with specific figures (35% emission reduction, €38 billion in 2022) replaces "generates revenue that can be used to fund renewable energy." The specific numbers make the claim verifiable and credible in a way that the original's vague statement cannot be.

Change 5: Make the revenue concrete

Rather than saying revenue funds "renewable energy initiatives and support for affected communities," the humanized version names the specific uses: transition support for coal workers in Poland, energy subsidies for French households. Specific examples are more persuasive than general categories.

Change 6: End with conflict, not resolution

"The hard part is not the economics. It is the politics." This ending is the opposite of the original's "represents a balanced approach." It leaves the reader with the actual problem rather than a tidied-up conclusion. This is more intellectually honest and more engaging.

The transformation above follows five consistent principles that apply to any academic or analytical essay paragraph:

• Open with your argument, not your context
• Acknowledge the tension or difficulty in your topic — AI avoids conflict, humans engage with it
• Replace every hedged general claim with a specific number, case, or example
• Use short sentences to vary rhythm, especially after long ones
• End sections with the real problem or implication, not a tidy summary

These are not tricks for fooling detectors — they are editorial principles that make academic writing better. The fact that they also reduce AI detection scores is a symptom of the quality improvement, not the goal.