Two types of people are reading this right now: those who feel calm and caught up on AI, and those who just felt their stomach drop a little because everyone in the lab seems to be casually fine-tuning models while they're still figuring out which button makes the chatbot go. If you're in that second group, take a breath. This one's for you.
OK, so think of it like the first day of a new class where it feels like everybody already read the textbook over summer. That sinking feeling has a name now. A Nature careers piece by Zhang-Ren Chen calls it AI FOMO - the fear of missing out, except the thing you're missing out on is "mastering AI" (Chen, 2026). The big, comforting punchline of the article is right there in the title: everyone is mastering AI except me - or are they?
Spoiler, gently delivered like I'm handing you a juice box: no, they're not. Most people are quietly bluffing too.
Why this feeling is so sneaky
Here's a simple way to picture it. Imagine a room where everyone is secretly nervous, but everyone is also really good at looking relaxed. You walk in, see all the calm faces, and think, "Wow, I'm the only nervous one." Meanwhile the person next to you is thinking the exact same thing about you. That's AI FOMO in a nutshell. It feeds on appearances.
The article points out that scientists are getting this from every direction. Conference talks, group chats, that one colleague who drops "I just had the model write my whole methods section" at lunch. It can feel like a race where the starting gun already fired and you were tying your shoes.
But think about how science actually moves. Slowly, with lots of stumbling. The folks who seem fluent usually got there by quietly fumbling through the same confusing tutorials you're avoiding because you're scared of fumbling. The fumbling is the learning. There's no shortcut elevator; everybody takes the stairs.
The part that should make you feel better
Now here's a neat little idea. You know how you don't need to understand exactly how an engine works to drive a car safely? Same deal here. The Nature piece makes the case that you don't have to become a machine-learning expert to use AI well in your work. You mostly need to know what these tools are good at, what they're bad at, and when to double-check them - because large language models will confidently make things up like your uncle explaining "how the stock market really works" at Thanksgiving.
That tendency even has a fancy name, "hallucination," and it's a well-documented thing researchers actively study and try to measure (Ji et al., 2023). So if you've ever caught a chatbot inventing a citation that does not exist, congratulations: you already understand one of the most important things about AI. That instinct to verify? That's the actual skill. The people who blindly trust the output are the ones to worry about, not you.
A gentle game plan
Let's break the scary mountain into small, friendly steps:
- Pick one tiny task. Summarizing a paper, cleaning up a figure caption, drafting an email. Just one.
- Try the tool on it. Then check the result like a careful teacher grading homework.
- Notice what it got wrong. That's you learning the tool's personality.
- Repeat next week. That's it. That's the whole secret.
If your "one tiny task" happens to involve wrangling PDFs - extracting text, splitting a giant supplement - browser-based tools like pdfb2.io can handle that privately without sending your unpublished work to some random cloud server. Small wins count.
The takeaway, celebrated properly
So, big gold star for you for even reading this far. The real message of Chen's article is kind, and I love it: the gap between you and the "AI masters" is mostly an illusion built out of confident faces and good lunchroom marketing. Curiosity beats panic every single time. You're allowed to learn this slowly. You're allowed to not know things yet. "Yet" is doing a lot of work in that sentence, and that's exactly the point.
Pretty reassuring, right?
References
- Chen, Z.-R. (2026). AI FOMO: everyone is mastering AI except me - or are they? Nature. DOI: 10.1038/d41586-026-01214-3. PMID: 42204328.
- Ji, Z., et al. (2023). Survey of Hallucination in Natural Language Generation. ACM Computing Surveys. DOI: 10.1145/3571730.
Disclaimer: This blog post is a simplified summary of published research for educational purposes. The accompanying illustration is artistic and does not depict actual model architectures, data, or experimental results. Always refer to the original paper for technical details.