top of page
Writer's pictureJoe Andrews

Speaking of: Will AI Kill Us All?

When I was in 4th grade, my class read this script from an episode of The Twilight Zone called "The Monsters Are Due On Maple Street." I have no idea why this was required 4th grade reading. The American literature curriculum is utterly unexplainable sometimes.

The episode is about a small suburban town that starts experiencing a bunch of weird, spontaneous electrical problems that none of the residents can really explain. A young boy suggests that maybe the electrical problems are being caused by a group of aliens living among them disguised as humans, and while everyone dismisses the boy at first, the residents slowly but surely start pointing fingers at each other and explaining why any little quirk a certain neighbor might have is proof that he or she is secretly an alien. The witch hunt quickly spirals out of control and the neighborhood descends into a full-fledged riot, collapsing under the weight of its own paranoia.

The episode ends with a slow camera pan out to two aliens staring heartlessly down upon the neighborhood. They reveal they were toying with the street's power to turn all the townspeople against each other and "let them destroy themselves," a strategy they've been washing, rinsing, and repeating in neighborhoods all across the country. As the scene fades to a star-filled outer space, a narrator says:

"The tools of conquest do not necessarily come with bombs and explosions and fallout. There are weapons that are simply thoughts, attitudes, prejudices ... to be found only in the minds of men. For the record, prejudices can kill ... and suspicion can destroy."

When we talk about AI destroying humanity, we normally talk about it in the Terminator sense of artificial intelligence growing more powerful than humans and killing us all or making us its slaves. That might still happen, and neither myself nor anyone working on AI knows whether it will or not. It's a complete uncertainty.

But what's not a complete uncertainty is that we are a deeply divided nation and world right now, and AI is indisputably going to put its thumb on the scale of that division. It will be a wedge that digs into every fracture in the human species and pries us further and further apart. It will generate misinformation that will make your uncle's Facebook wall look like child's play. It will give hackers and other bad actors the type of cybercrime toolkit Putin has wet dreams about. It will give autonomous weapons the merciless, boundless, and senseless killing power that make the Navy SEALs look like toy soldiers by comparison. It will bring out the worst of humanity at a scale previously unthinkable and at a pace practically uncontrollable.

I'm an optimist by nature. The high majority of my time thinking about AI is spent thinking about the opportunities for good in education and business planning and brainstorming and the creative arts. But the only way to properly prepare for the downside is to look it in the eyes and call it by its name. I don't know if the Terminator future of AI overlords will ever happen, and quite frankly I don't care. Because what I do know is that AI will make us kill each other long before it ever gets close to killing us. We'll die by our own hands before we even find out if we'd survive in its. I urge anyone trying to protect humanity from the dangers of AI to worry less about The Terminator and more about the aliens hovering over Maple Street.

Commentaires


bottom of page