Academic English revision

AI’s Default to 'He' Is Sabotaging Your Social Science Draft

AI defaults to singular 'he' because its training data is decades old, but your paper shouldn’t. Singular 'they' is now the standard in social science journals—use it deliberately, not blindly.

You just generated a social psychology discussion draft, and the first thing you notice is an ocean of 'he,' 'his,' and 'him.' Your advisor’s margin note says two words: "dated." You know it’s wrong, but you don’t know how to fix it without making the whole thing sound like a high school essay. This isn’t an English problem. It’s a stale data problem, and your draft is paying for it.

Here’s the truth nobody tells you: AI favors generic 'he' because its training corpus is heavy with pre-2000 academic texts, when singular 'he' was the default for gender-neutral reference. Your paper doesn’t need to echo 1985. When you write “a researcher” and follow it with “he,” you’re not demonstrating authority—you’re showing your writing hasn’t kept pace with disciplinary evolution. This isn’t politics. It’s academic currency.

The most common failure: AI writes “a student who submits late... he receives...,” and you let it stand. The fix isn’t a word swap; it’s a perspective shift. ❌ A researcher who collects data must ensure that he uses valid instruments. ✅ A researcher who collects data must ensure that they use valid instruments. Note: 'they' takes the plural verb form—'use,' not 'uses.' This is where AI trips itself up, so don’t just ask it to “check.” Ask it to reconstruct.

Trap number two: AI loves writing as if the reader is in the room. It generates “When you read a text, you often...” and calls it engaging. In a journal article about cognitive mechanisms, it reads like a blog post talking to you. ❌ When you encounter uncertainty, you tend to rely on heuristics. ✅ Individuals encountering uncertainty tend to rely on heuristics. Or tighter: Under conditions of uncertainty, reliance on heuristics is common. Elevate 'you' to a concrete noun—individuals, participants, consumers—and the register snaps back to academic. I don’t fix this manually. I dump the draft into easydue and let it flag every generic-'you' instance with three rewrite options. The time I save goes to sharpening my argument, not hunting pronouns.

Trap three is the sneakiest: mixed reference. AI writes 'they' in one sentence, then snaps back to 'his' two sentences later because its attention window linked 'participant' to a specific example. ❌ The participant shared that they felt anxious. His response was delayed. ✅ The participant shared that they felt anxious. Their response was delayed. Or more precise: The participant reported feeling anxiety; their response latency increased by 1.2 seconds. The fix isn’t memorizing which pronoun belongs to which noun. It’s building a scanning habit: after each paragraph, search for he/his/him and check every instance. AI generates the first draft; consistency is your job.

One point almost everyone misses: 'they' isn’t a universal fix-all. When your text refers to a specific gender group, forcing 'they' reads as clumsy. ❌ Nurses, who are predominantly women, face burnout. ✅ Female nurses report higher rates of occupational burnout than male counterparts. Be specific when the referent is specific; be generic only when it truly is. AI stumbles here because its attention mechanism struggles to track whether a passage discusses a group or an individual. You don’t need fancier vocabulary. You need sharper referential strategy.

Steal this rule of thumb: In social science writing, use singular 'they' for unspecified individuals. Use specific nouns or adjectives for defined gender groups. Upgrade generic 'you' to concrete agents like individuals, participants, or consumers. Those three rules cover 95% of the mess. Next time you run a draft, scan your pronoun density in easydue. Watch those stray 'he’s' disappear one by one, and you’ll feel your writing tighten into something a journal editor won’t flinch at.