Most people will agree that AI is an impressive writer. In seconds, it spits out text that stuns even skeptics. But there’s a caveat. Two, to be precise.

1. The AI voice. When we write, our essays exude our personalities; they reveal our idiosyncrasies and quirks. We have opinions about most things and verbalize them in unique ways. Our thoughts have an emotional or rational appeal in a ratio that reflects our mood on a given day. AI does not have a mood, although you can fake it. But it is not you.

2. The AI bias. AI’s opinions are deeply rooted in the biases of its programmers. It may seem a bit liberal or conservative, but it will never take a stand on polarizing issues. Instead, it will always take the apologetic “as a large language model” line.

So we must decide to what extent the AI’s written output reflects our own opinions, beliefs, and word choices. Is the output formal, informal, or a little bit of everything? Editing an AI goes far beyond checking syntax (which is usually not required).

As most users take AI-generated text at face value or merely review the content for minor improvements, we are moving toward a future where AI becomes a cognitive influencer. Its ubiquitous presence ranges from silly social media posts about dogs to silly social media posts about elections. AI can incite people and fuel endless discussions with its encyclopedic knowledge that may serve hidden purposes.

So let’s relinquish the role of clueless editors, take the reins of our own destiny, and ride toward a future where humans still hold the wheel. We stand at a crossroads where we either reclaim responsibility for our writing or delegate it to synthetic beings whose decisions may bring unpleasant consequences. Pick up your pen!