I've been worried about x-risk for a very long time. But for some reason AI x-risk has never felt intuitive to me.

On an intellectual level AI risk makes sense to me, but I would like to be able to visualize what a few plausible AI take-over scenarios could concretely look like. I would really appreciate if anyone could share books and movies with realistic AI elements and great storytelling that helped you to connect emotionally to the scariness of AI x-risk. 

Alternatively, I would appreciate recommendations for any content, posts, ideas, revelations, experiences, "a-ha" or "0h-shit" moments, etc. - anything that helped you realize how seriously scary AI is and motivated you to work on it.

Thank you for the advice!




New Answer
New Comment

7 Answers sorted by

My favorite is probably the movie Colossus: the Forbin Project. For this, would also weakly recommend the first section of Life 3.0. 

Ex Machina does a great job of demonstrating the AI boxing experiment.

Besides the typical (over-dramatised) killer robot scenarios, I would add the perspective of looking for infrastructure breakdowns or societal chaos. Imagine a book like Blackout with a disruptive, national blackout but that was caused by powerful intelligent (control) systems. 
Or a movie like Don't Look Up where AI is used to or actively spreads misinformation that severely impact public opinion and taking effective action. 
In movies and books it might commonly be portrayed (in part) as human failure; but it could have been the result of correlated automation failure or goal mis-specification or a power-seeking AI system. 

The consequences that individual humans and society at large suffer from critical infrastructure breaking down can be quite realistic and visceral. 

I think the series Next did a pretty good job of making me scared. It's not an amazing production in itself, but worth watching.

What this post and answer shows me… is that we probably need more material? If we want to ‘scare people’ about AI and make more people aware of it as an issue.

We’ve got lots of great (and dense) YouTube and blog content on the technical aspects. We could do with more more “Holy Shit Imagine an AI That Does This!!” type-content.

Some examples not mentioned already, (and of course YMMV for each):

That Alien Message by Yudkowsky is very short, and might get at the intuition

It Looks Like You’re Trying To Take Over The World by Gwern is full of annonated links, if you want to get very technical and concrete about things

For a full-length book, I think that the Crystal Trilogy by Max Harms does a very good job, and would recommend if you like similar sci-fi

Finally, I would actually recommend the film Upgrade. While nowhere near as good a film as Ex-Machina, it's a good illustration that the AI risk is about malicious software agents, and not "evil robots" (i.e. Terminator) which I still think is the dominant public image when prompted about AI risk. Also, really cool fight choreography.

For me, the easiest to imagine model of how an AI takeover could look like has been depicted in Black Mirror: Shut Up and Dance (the episodes are fully independent stories). It's probably just meant to show scary things humans can do with current technology, but such schemes could be trivial for a superintelligence with future technology.