Will AI Save Us, or Just Make Us Better at Being Bad?

We’ve always pictured aliens as the threat, mysterious, advanced civilisations arriving unannounced with weapons we can’t comprehend. But here’s the uncomfortable twist: when first contact finally happens, we might be the ones they’ve been warned about.

And if AI keeps evolving the way it is, that warning might come with a flashing red light.

The case for AI saving us

In theory, AI could finally help us outgrow our toddler-with-a-grenade phase as a civilisation.

• No ego, no grudges, no “he looked at me funny” diplomacy, just cool, calculated problem-solving.

• AI-led conflict modelling could identify dangerous escalations and stop them before they start.

• Global resource management algorithms could reduce scarcity, ending the centuries-old tradition of fighting over oil, water, and whichever lump of rock we think is worth planting a flag on.

Imagine a future where AI is the calm adult in the room, patiently suggesting “Maybe don’t blow up the planet today” while the rest of us argue about who gets to sit in the captain’s chair on the first Mars flight.

In this optimistic timeline, AI isn’t just our safety net, it’s the pilot steering us into a more peaceful, cooperative era. Space exploration becomes a joint venture. Interstellar diplomacy replaces interstellar conquest. We show up in the galaxy as the friendly neighbour offering to help fix your star-drive, not the one stealing your lawn furniture.

The case for AI turning us into galactic nightmares

Then there’s the other version, the one human history suggests we’re far more likely to pick.

AI doesn’t have to fix us. It could just make us better at being awful.

• Designing weapons in seconds that make current nukes look quaint.

• Mapping alien planets in forensic detail before we’ve even said hello, complete with resource extraction plans.

• Analysing hypothetical alien defences and suggesting the most “efficient” invasion strategy.

With AI as our co-pilot, we wouldn’t need centuries of trial-and-error conquest. We could scale up our worst traits immediately, greed, paranoia, opportunism, and project them across the stars with laser precision.

If that’s the road we take, our first contact moment won’t be a handshake. It’ll be a galactic version of a SWAT raid. And in that case, we become the cautionary tale alien parents tell their kids to get them to eat their nutrient paste.

The irony no one talks about

We’ve built entire genres of movies, books, and paranoid late-night podcasts around the idea of hostile alien invasions. “What if they’re dangerous?” we ask, conveniently ignoring the historical record that says we’re the species with a proven habit of turning up uninvited and taking whatever we want.

From the perspective of any peaceful civilisation out there, Earth might already look like the suspicious neighbour with too many locks on the door and an unusually high number of “accidental” fires.

Now throw advanced AI into that picture. Suddenly, our ability to act on those aggressive impulses scales exponentially. We wouldn’t just be a threat to our nearest neighbours, we’d be a threat to civilisations we haven’t even met yet.

And that’s the part that should make us squirm: if the galaxy is full of watchers, they may be less worried about protecting us from danger… and more worried about protecting everyone else from us.

The restraining order scenario

Imagine it: humanity finally achieves faster-than-light travel, bursts onto the interstellar stage… and the first thing we get is a polite but firm request to turn around and go home.

If we’re lucky, it’s just a written notice. If we’re not, it’s a blockade.
Either way, the irony is complete, after centuries of fearing hostile aliens, we discover we are the hostile aliens.

The fork in the road

AI could still steer us toward the better path. It could help us manage our instincts, solve problems without violence, and arrive in space as partners rather than predators.

But if we build it in our image, with all our biases, rivalries, and hunger for power baked in, then we’ve just created the ultimate enabler.

The question isn’t just whether AI will save us from ourselves.
It’s whether it will save them from us.