Warning! This post deals with ideas about death and the afterlife. If that sort of thing makes you uncomfortable buckle up or hit the eject button.

I’ve been having some really strange daydreams recently about possible intersections in our time stream of pop culture, advances in machine learning and the afterlife. Mainly about how people will cope with and explore the facts of deeply complex artificial reproductions and cheap always-on cloud services.

What got me thinking about this was the wealth of knowledge and experience that has been building in the machine learning world around language and speech analysis, our absolute addiction to information technology and our occasional quest for deeper meaning in a world that is largely out of our control. (I think the afterlife stuff got mixed in because someone was watching Game of Thrones in the background.XD)

As for the machine learning aspects, I’m thinking about things going all the way back to likes of ELIZA, PARRY, and A.L.I.C.E. up to much newer explorations like “What Would I Say” and Reddit’s r/SubredditSimulator.

“What Would I Say” is a bot that will create Facebook posts for you based on Markov chains trained from your old material. The SubredditSimulator also has Markov chains at its heart, but I can’t do any greater justice to describing it than their own words

… this is a fully-automated subreddit that generates random submissions and comments using markov chains… with each bot account creating text based on comments from a different subreddit.

Both of these bots operate by harvesting comments from histories that are contained on the host platforms. In the case of the Facebook bot, it analyzes your own history, but the Reddit bot will look across the comments from all the users of that site. Both are intended to create realistic facsimiles of content that might actually come from a human. Of course, in the case of the latter I feel the intent is less focused and more a product of the community culture at Reddit, but I find it to be a curious experiment.

Then you get to something really odd, and slightly scary, with this headline from last year “Postdoc develops Twitterbot that uses AI to sound like Donald Trump”. That really started to make me wonder about other applications for this type of technology and where we might be going culturally.

Could we start to build digital models of ourselves that we become comfortable enough with to the point where we would consider turning them on in the event of an early demise? Might our relatives do something like this posthumously just to hear something akin to our voices one more time?

Zombie Twitterbot speaks

At some level this would just be another form of memorial, like choosing if you will have a funeral or what prose you would like on a headstone. Not everyone has the opportunity, the resources, or the desire to undertake these types of memorials, but some do.

It got me questioning how long would it be before we saw someone trying to start a business offering this type of service. I mean, hey machine learning is hot, cloud services are hot, startups are fun, you do the math. Maybe this has already been proposed, I’m not sure if we’ve crossed that rubicon yet. I’d love to hear about it if anyone is pushing this boundary in cloud-aware funerary memorial as-a-service offerings. Then again maybe my mind isn’t quite ready for that yet.

I had some more thoughts along this line, but perhaps those fevered dreams are best left for another flight of fancy. Thanks for listening and joining me on this ride, I hope you had a good time =)