You Live in the Deepfake Era. Be the One Who Checks Twice.

This is our reading of Bill Gates's 2023 essay The risks of AI are real but manageable (Gates Notes). Read the original →

First, accept one thing.

You are the first generation in history that has to default to "this video I am watching might be fake." A speech by a politician, a clip of a celebrity, even a voice note from a friend: all of them could be AI-generated, and the generation is getting better. In 2023 Gates wrote: "In an era of deepfakes and misinformation that can be tailored to you individually, the ability to tell what is true from what is not becomes an essential life skill." Three years on, that is not a warning. It is what your phone hands you every morning.

This sounds frightening. But in the same essay Gates said something else, and he had good reason to say it.

"The best reason to believe that we can manage the risks is that we have done it before."

What does he mean? Think about the wave you did not catch but your parents did: in the late 1990s, scammers sent emails pretending to be Nigerian princes, offering to share their fortune if you helped move some money. Many people fell for it. Slowly, people learned. Today when you see a scam email, you do not even have to stop and think. That reflex was built by time. For deepfakes, the same muscle has to grow, but faster.

Gates uses another analogy that is even more useful.

When calculators arrived in classrooms, one group of math teachers panicked. Children would forget how to do arithmetic. Another group turned around and taught the reasoning behind arithmetic: why carrying works, why the decimal is placed where it is, what addition really means in a shop. The first group lost to the calculator. The second group produced students who were better than any calculator, because the students now understood.

You are standing at the same fork with AI. AI will write your assignments. That is a fact. The question is whether you will get worse because of it.

Gates cites an English teacher, Cherie Shields, who wrote a piece for Education Week in early 2023. She did not ban ChatGPT. She built her lessons around it: teaching students how to prompt it, how to read the draft it produced, how to spot what was wrong, how to rewrite it into something better. Her students did not learn to write less. They learned to write more. Because for her, AI was a whetstone, not a ghost-writer.

You can do the same.

Every time an AI writes something for you, do not submit it. Do three things first.

One: read it out loud. You will hear the hollow sentences, the elegant but empty phrases. Those are the AI's tics, not yours.

Two: pick one concrete claim and go verify it. AI said "research shows"? Which research, by whom, published where? If you cannot find it, delete it. If you can, now you have actually learned something.

Three: rewrite the most important paragraph in your own words. The result will read more like you, though less polished. That roughness is your signature.

Ten minutes, total. But what those minutes do is turn AI from "the thing that gets me through one more assignment" into "the thing that trains my head one more time."

For images and videos, the rule is simpler.

When a piece of content makes your emotions spike, especially rage, especially wonder, especially fear, pause. Ask three questions.

Where did this come from? A first-hand source, or a screenshot of a screenshot?

Who benefits if I believe this? A traffic engine? A company? A political camp?

If I wanted to prove it fake, what would I look up? Is there a reverse image search? Is there original footage?

Practice those three questions until they become reflex. Then you become the person in the group chat who first says, "wait, this might be fake." In three years, that person is who everyone else trusts.

One last thing.

Your classmates are already sorting into two piles. One uses AI to finish assignments, write captions, write apology texts, and slowly loses the ability to think a problem through. The other uses AI as a whetstone, and gets faster, clearer, sharper at telling true from false. Ten years from now, those two piles will not stand on the same starting line. Which one you land in is your call, and only yours.

### Questions to sit with

1. The last time you shared a video or an image that "hit you hard," did you stop to check the source? If not, how quickly did you pass it on?

2. Who in your circle is the person who most often says, "wait, this might be fake"? What do they bring to the group?

3. If a year from now your teacher tells the whole class, "you can use AI on this assignment," which version of you do you want to be that day: the one whose AI finished the work, or the one who used AI to get better at doing the work?