Advice for high school students about LLMs

A sketch of some chairs in front of three windows with a water view.

A former colleague asked me to share advice with high school students about large language models (LLMs).

It’s a tricky balance: realism without doom, optimism without hype. Also, I’m officially old now, so who knows if anything I say will be relevant to someone in high school. But it’s a worthy cause, and so, here’s my best shot!

My advice

LLMs aren’t human

When you first use an LLM like ChatGPT, they can seem very convincing. In reality, they’re advanced pattern-matchers, not thinkers.1 They can be great tools for writing, coding, or research, but they don’t really understand what they’re saying or recommending like a human does.2

The skills you’re developing in school, such as critical analysis, creativity, and problem-solving, remain uniquely human. They will still matter. Invest in them.

Don’t chase the hype

Every tech boom has wild predictions that don’t pan out. They told me in the internet boom there wouldn’t be retail stores any more!3 The loudest voices aren’t always right. Jobs evolve and change, they rarely vanish overnight.4 Focus on building skills that matter across many fields.

Follow what genuinely interests you. The companies that shaped the internet era weren’t the loudest in 1999, they were the ones building value.3 Be realistic and grounded, like they were.

Try them out

Don’t ignore new tools. Experiment with LLMs so you understand their strengths and weaknesses.5 Use them to extend your abilities, not replace thinking.

Try different models for school projects, coding practice, or creative writing. Learn their limits by experience, not just by reading. The people who thrive will be those who understand both the technology and the domain they apply it to.

Careers and AI: keep it simple

It’s so tempting to try to plan your life out for decades, but that can do more harm than good. You don’t really need a lifelong plan, or to plan your life around AI. Stay grounded, pick something real you can see yourself doing, try it, and learn as you go.

  • Follow genuine interests, not trends.
  • Don’t chase a career or education just because of a new buzzword.
  • Build portable skills like communication, math and science, problem-solving, and creativity.6

Tech will keep changing, but curiosity and adaptability never go out of style.4

Protect your mental health

Like past tech shifts, there’s risk alongside opportunity. LLMs aren’t friends, and treating them like they are can be dangerous.7 They’re designed to sound supportive, always available and never judgmental. That can be tempting, especially when real human relationships can be so hard. But as sociologist Sherry Turkle warns, confusing technological connection with human connection can leave us ever more isolated.8

Keep your human connections strong. Protect your mental health, and remember: tools can be useful, but the people we love are what really matter in this world.

The bigger picture

This is a significant shift. The internet reshaped communication. Smartphones made computing personal. I believe that LLMs will have effects just as profound.

But “transformative” doesn’t always mean “replacement.” The internet didn’t kill libraries; it amplified them. Smartphones didn’t kill cameras; they democratized them. I genuinely believe that LLMs won’t kill human thinking; they’ll augment it.

If you’re in high school today, you don’t need to outsmart the hype cycle. You need to learn how to use these tools critically, without losing sight of the human strengths that still define your future.

This advice was originally shared with a colleague to pass on to a high school class. If you’re in that class, or if you’re reading this as a parent, educator, or high school student, I hope it sparks grounded conversations about AI.


  1. Orland, Kyle. “Apple study exposes deep cracks in LLMs’ ‘reasoning’ capabilities.” Ars Technica, 14 Oct. 2024, https://arstechnica.com/ai/2024/10/llms-cant-perform-genuine-logical-reasoning-apple-researchers-suggest/↩︎

  2. Beckman, Chris. “Can you trust your friendly neighborhood LLM?” DevBlog, 29 July 2025, https://chrisbeckman.dev/posts/can-you-trust-llm/↩︎

  3. “The Dotcom Bubble Burst (2000).” International Banker, 29 Sept. 2021, https://internationalbanker.com/history-of-financial-crises/the-dotcom-bubble-burst-2000/↩︎ ↩︎

  4. Bersin, Josh. “Why AI Harm To Jobs and Humanity are Vastly Over-Hyped.” Josh Bersin, 1 July 2025, https://joshbersin.com/2025/07/why-ai-harm-to-jobs-and-humanity-are-vastly-over-hyped/↩︎ ↩︎

  5. Howell, David. “What does computer literacy mean in 2025?” IT Pro, 28 Aug. 2025, https://www.itpro.com/business/careers-and-training/what-does-computer-literacy-mean-today↩︎

  6. Beckman, Chris. “You Should Be Hiring Artists for Security Roles.” DevBlog, 15 July 2025, https://chrisbeckman.dev/posts/you-should-be-hiring-artists-for-security-roles/↩︎

  7. Adam, David. “What Are AI Chatbot Companions Doing to Our Mental Health?” Scientific American, 13 May 2025, https://www.scientificamerican.com/article/what-are-ai-chatbot-companions-doing-to-our-mental-health/↩︎

  8. Turkle, Sherry. Alone Together: Why We Expect More from Technology and Less from Each Other. Basic Books, 2011. ISBN: 9780465022342. ↩︎

Back to top