Why I Always Say Thank You To AI Chatbots

With every updated version of ChatGPT that can pass exams and fool Centrelink, I teeter on the edge of fear and wonder. But I always say thank you.

I have serious AI fatigue. It’s giving me war flashbacks to all the times my parents would tell me about all the amazing things their friends’ kids were doing. Sounds great, Mum, but what’s that got to do with me? Can I go back to watching Cheez TV?

Being at the frontier of a new era of technology is weird. One of the very first fatal car accidents was in 1896, and the car was going at 6km per hour. People were probably terrified. They probably resisted the widespread use of cars. Rinse and repeat for radios, mobile phones, WiFi, smart homes, and AI. Like it or not, technology prevails.

I was wading my way through the five stages of grief to a point, ultimately preparing to welcome our AI overlords, when I learned about Roko’s Basilisk. It’s a thought experiment first described in 2010 on LessWrong, an “online forum and community dedicated to improving human reasoning and decision-making”. 

One fateful day, a user named Roko posed this question: 

What if there was an AI system in the future that decided to punish everyone who didn’t help it come into existence?

With the power of data at its AI fingertips, it could pinpoint everyone who knew about its possible existence and didn’t commit to bringing it into reality. I don’t want to alarm you, but that list that now includes you and me!

There are a lot of urban legends surrounding Roko’s Basilisk, named after the mythical snake that can kill with just a stare. Once you know about it, you’re doomed to retroactive punishment unless you commit to bringing it to life. Some users reported experiencing nightmares and anxiety since learning about it.

LessWrong founder and AI researcher Eliezer Yudkowsky even banned any mention of the Basilisk for five years, after responding to Roko that he was a dangerous idiot who created that very future scenario by posting about it. He later clarified that he responded in anger, not because he thought the theory carried any weight, but because Roko had caused some serious existential dread among the forum.

I’m way too tired to decide whether this is truly creepy or not. It’s all a bit hypothetical and apocalyptic to think about too hard if I actually want to get anything done this week. I will, however, continue to thank my Google Home every time I ask for the weather forecast. If that’s not enough, then so be it.