The real risk of AI isn't the machine - it's us

17 November 2025·Billy Cheung

Source: Pexels

Artificial Intelligence is everywhere right now - in headlines, strategy decks, and late-night conversations about whether we'll all be replaced by robots.

But, before we start declaring ourselves "AI experts" overnight or panicking about AI eliminating humanity, it's worth pausing to ask: what kind of "intelligence" are we really dealing with?

Let's take a step back and look clearly - what AI really is, what it isn't, and how to approach it responsibly, as it was always meant to be: a tool.

What exactly is AI?

Despite all the hype, AI isn't some all-knowing digital brain, nor is it dark magic. It's not even particularly intelligent - at least not in the human sense.

What we call "intelligence" is really an enormous database performing lightning-fast pattern matching and returning the closest result - all wrapped in a chat interface that enables us to interact with pattern calculations through language, giving the illusion that we're "chatting" to someone conscious on the other end.
In other words, AI doesn't think or understand; it estimates.

That's why it can sound like a genius one moment and hopelessly wrong the next. It's artificial, not conscious. And because it learns from us, it inherits our blind spots, biases, and mistakes.

It's essentially like the internet: full of information, yet only as reliable as the sources behind it. The same principle applies - we shouldn't believe everything provided on the screen. We'll need that same healthy scepticism when dealing with AI outputs.

Can AI replace humans?

The short answer: no.
The longer one: it depends what "replace" means.

AI can automate tasks, speed up analysis, and even complete things faster than we can. But, accountability is still on the humans.

IBM once reminded its teams:

Material from IBM back in 1979.
Reposted from Medium

The same principle still stands today, and it is even more relevant with AI.
It's like the oven in a kitchen. The heat cooks food for the chef, but it's the chef who decides whether the food is properly cooked.

A lesson in relying too much on AI

In July 2025, the entire live database of a software company was erased - just because of a software engineer's experiment with an AI-assisted "vibe coding" tool.

It happened because too much privilege was granted to the AI tool, and human gatekeping wasn't carefully enforced.
As a result, machines act, and humans answer for the outcome.

So yes, AI might be far more efficient than us. Yet utimately, it's still humans who are held accountable.
Using the same oven-and-chef analogy - when people are served undercooked food, the blame still falls on the chef who failed to check it properly before taking it out of the oven.

So, is AI bad?

It's not really a question of whether AI is good or bad.
It's a tool - and like any tool, it can be used for a good cause, as well as for bad intent.

Scammers use AI to impersonate. Medical workers use it to fight deceases. The technology itself is neutral - it's the intent that define the outcome.

What actually matters is, on top of using it ethnically - we need to use it responsibly, and we should keep learning.

A calculator can sum up numbers quickly, but we still need to understand the maths behind the numbers. Otherwise, it's impossible to judge if mistakes are made using the calculator.
The same principle applies for AI.
It can identify potential patterns beyond human detection and generate seemingly-convincing answers. But, without critical interrogation and foundational understanding of the subject, it is simply ignorance and blind trust on something error-prone, resulting in the broadcast of fasle information.

Think of Google. When it first appeared, it didn't make anyone smarter on its own.
But for those willing to keep learning, it opened access to knowledge they'd never have reached before.

AI works the same way. In the hands of the curious, it's an amplifier. In the hands of the careless, it's a megaphone for mistakes.

AI isn't rewriting what it means to be human - it's reminding us to stay responsible, curious, and resilient.
The real challenge isn't about machines becoming more smarter, but about humans becoming wiser users.