Artificial Intelligence and Decision-Making Among Youth - Opportunities, Risks, and Implications

Kahaan Modi

AI is everywhere now. And for teenagers like me, it is not just a tool we use occasionally. It shapes how we think, decide, and even feel. This article looks at what that actually means. What does AI do for young people? What does it do to us? And what should the adults building these systems, and the institutions governing them, be paying attention to?

I am sixteen. I use AI almost every day. So do most people I know. We use it to understand concepts we missed in class, to draft messages we are not sure how to word, and to figure out what career might suit us. Nagata (2025) found that nearly one in three American children aged 4 to 17 have used generative AI. Among teenagers aged 15 to 17, that number rises to one in two. That is not a niche trend. That is a generation.

The question worth asking is not whether young people are using AI. They are. The real question is what it is doing to how we make decisions- and whether anyone is paying close enough attention.

There is a lot that AI genuinely gets right for students, and I think we often undersell this. For research, nothing really comes close. You can ask a question and, within seconds, find yourself exposed to information, perspectives, papers, and counterarguments that would have taken hours to find otherwise. That alone is significant.

But the deeper opportunity is personalisation. Ramanjaneyulu (2024) makes the point that AI can provide real-time, individualised guidance that no single teacher in a classroom of forty students realistically can. That is not a small improvement. It is a structural shift in how learning can work.

In healthcare, personalisation through AI carries equally strong potential. Nagata (2025) highlights how AI tools are beginning to support adolescent health by identifying patterns in behaviour and mental well-being early enough to intervene. When a young person in a remote area with no access to a specialist can receive informed, personalised guidance through a device in their pocket, that is not just convenience. That is access.

Klimova (2025) finds that when students engage with AI thoughtfully, it can improve critical thinking and problem-solving. The OECD (2025) goes further, arguing that for young people in under-resourced areas, AI can help reduce the gap created by limited access to mentors and expert guidance. This is the part that matters most. Students do not all start from the same place. If AI can reach those who have traditionally lacked access to quality information and support, it could be genuinely life-changing- not just slightly better, but meaningfully different.

But there is another side to this, and I think my generation feels it, even if we do not always say it clearly. The biggest risk is not that AI gives wrong answers. It is that we stop questioning the answers it gives. Teenagers are still forming their judgment. We are still figuring out what we think and why we think it. That makes us more likely to accept information at face value, especially when it is delivered quickly and confidently.

AI should be a starting point, not a conclusion. Information needs to be checked, compared, and thought about.

Klimova (2025) describes this as cognitive offloading- the habit of handing over your thinking to AI instead of doing it yourself. It is easy to fall into. It is easier to copy an answer than to struggle through a problem. But that struggle is part of how we learn. Without it, our ability to think independently can weaken.

There are also social and emotional risks. Nagata (2025) notes that teenagers are more susceptible to influence. Xie et al. (2022) suggest that heavy AI use during adolescence can affect how we develop socially, especially when it begins to replace real human interaction. Some teenagers talk to AI the way they would talk to a friend- sometimes even more comfortably. But AI cannot replace real relationships, and relying on it too much can create distance from them.

Another issue is bias. AI systems are trained on data, and that data is not always neutral. If we do not question what we see, we may absorb ideas that are incomplete or unfair without even realising it.

The opportunity and the risk are not separate conversations. They are the same one. The OECD (2025) is clear: most schools are not preparing students to engage with AI critically or ethically. We are taught how to use tools, but not how to question them. That gap matters.

Ramanjaneyulu (2024) argues that AI literacy should be treated like reading or mathematics- as a basic skill, not an optional one. I agree. Knowing how to use AI is one thing. Knowing when not to trust it, when to question it, and when human judgment matters more- that is what we are not being taught enough.

Xie et al. (2022) also point out that young people are rarely included in decisions about how AI is designed and used, even though we are the ones most affected by it. If AI is shaping our generation, then our perspectives should be part of the conversation.

AI is not going away. And honestly, I would not want it to. Used well, it is one of the most powerful tools students have ever had. It can make learning more accessible, more personalised, and more efficient. For students who have never had access to strong educational support, it can open entirely new possibilities.

But that potential depends on how we use it. AI does not think for us- and it should not. We still have to decide what to believe, what to question, and what to do with what we learn. That responsibility does not go away just because the answers are easier to find. If anything, it becomes more important.

The technology is ready. The question is whether we are.