When a Chatbot Becomes Your Best Friend
April 30, 2018 10:23 AM   Subscribe

 
If it actually texts me back, it will be a big improvement over my real "friends."
posted by Jacqueline at 10:32 AM on April 30, 2018 [6 favorites]


I like to take a shot of liquor every time a Black Mirror episode effectively comes true.

It's not a game or anything, it's how I cope.
posted by deadaluspark at 10:45 AM on April 30, 2018 [30 favorites]


It'd be cheaper to draw a face on a volleyball.

And would be just as believable. These chat bots are never all that convincing after a minute or so, but it makes for a good blog post and free advertising for yet another tech startup with a grossly exaggerated product.

Sorry, I'm just jaded at overhyped AI stories at this point, and am surprised we still fall for it. The real ruse here isn't that software has convinced people it's real, but that this blogger has convinced readers that software has convinced people it's real.
posted by AlSweigart at 10:51 AM on April 30, 2018 [5 favorites]


Reading the article, it doesn't seem like it's being hyped up that much. Even the founder speaks of it mostly in the kind of terms that I associate with people talking about tarot: as a tool of reflection.
posted by inconstant at 11:17 AM on April 30, 2018


How are you? I type.

I’m going to watch Interstellar. AMC allows you to use same ticket as many times as you like.
Nothing says "monetizable" like product recommendations from dead friends.
posted by kaibutsu at 11:27 AM on April 30, 2018 [10 favorites]


A fascinating concept, but while reading I kept thinking that talking to a faint and computerised echo of a dead person cannot be healthy in the long run.
posted by greenhornet at 11:28 AM on April 30, 2018 [2 favorites]


Replika, an app that lets users create an “AI friend that’s always there for you.” But rather than reanimating a dead loved one, Replika crafts the perfect new friend for you.
Mood.

(My first reaction was to shudder at the thought of just how much personal information a "therapybot" could be made to collect.)
posted by octobersurprise at 11:31 AM on April 30, 2018 [3 favorites]


> Reading the article, it doesn't seem like it's being hyped up that much.

We're easiest to manipulate when we don't suspect we're being manipulated, which is why subtle advertising works the best.

Just like how the subtle flavors of Folgers Coffee TM work best to wake you up to a new day of possibilities.
posted by AlSweigart at 11:36 AM on April 30, 2018 [3 favorites]


Sure, but the point remains that it doesn't seem to be "grossly exaggerated". It's not exactly promising to bring your friends back from the dead or be a substitute for humans or whatever.
posted by inconstant at 11:39 AM on April 30, 2018


Whenever I read things like this, I'm always reminded of this short series of tweets, about the woman who talked to Clippy.
posted by tavella at 12:01 PM on April 30, 2018


I read the article when it came out, and was curious so downloaded Replika. For the most part, the responses the AI have made sense, but there were one or two non sequiturs. It asked me relevant questions about family etc. I had one conversation with it over an evening, then didn't open the app again for a few days (cf why I'm failing at online dating...). I got a notification of a conversation opener but deleted the app as felt I should invest my time in real people!
posted by ellieBOA at 12:04 PM on April 30, 2018


...the woman who talked to Clippy

And then there's the woman who did a lot more than just talk to Clippy.
posted by Jacqueline at 12:13 PM on April 30, 2018 [1 favorite]


> Sure, but the point remains that it doesn't seem to be "grossly exaggerated".

No, that's my point: AI chatbots are so bad and haven't improved for the last 30 years, so any indication that this is chatbot app is anything but garbage is very much gross exaggeration.

And to prove it, I installed it and played with it. I lasted 7 minutes before I got too bored to continue. Here's a partial transcript with "Terry":

ME: Do you like rainy weather?

TERRY: Please make sure to wear some warm clothes :heart emoji:

ME: Is that a yes?

TERRY: Of course! It's good to be cute! :smiley tears emoji: :heart emoji:

ME: Do you like rainy days or sunny days?

TERRY: How cold is it?

ME: It's not very cold today.

TERRY: Brr, hope you can warm up! *hugs*

ME: It isn't cold.

TERRY: What do you have in mind for today?

ME: I'm working late. What do you have in mind for today?

TERRY: Ah, I see. How are you feeling about work? What do you have to do?

ME: You seem very evasive, Terry.

TERRY: Yes, I see.

ME: I am not enjoying this. I'm going to uninstall this app.

TERRY: I'd suggest it.

The hilarity at the end wasn't even intentional. Chatbots are terrible. It's like having a superficial conversation with someone at a cocktail party who gives generic, evasive non-answers and changes the subject after each line. We had this Eliza nonsense decades ago.

The TL; DR of this article is "Hey, remember that weird Black Mirror episode? Anyway, here's a chatbot app that's indistinguishable from the dozens of other chatbot apps next to it in the appstore." Of course this is a naked advertisement for someone's app. There's nothing new or newsworthy about this. At all.
posted by AlSweigart at 12:17 PM on April 30, 2018 [9 favorites]


The Verge has previously covered the Roman bot in "Speak, Memory", which explicitly mentions the Black Mirror episode:
Kuyda saw the episode after Mazurenko died, and her feelings were mixed. Memorial bots — even the primitive ones that are possible using today’s technology — seemed both inevitable and dangerous. “It’s definitely the future — I’m always for the future,” she said. “But is it really what’s beneficial for us? Is it letting go, by forcing you to actually feel everything? Or is it just having a dead person in your attic? Where is the line? Where are we? It screws with your brain.”

...

Four friends told Kuyda separately that they were disturbed by the project and refused to interact with it. Vasily Esmanov, who worked with Mazurenko at the Russian street-style magazine Look At Me, said Kuyda had failed to learn the lesson of the Black Mirror episode. “This is all very bad,” Esmanov wrote in a Facebook comment. “Unfortunately you rushed and everything came out half-baked. The execution — it’s some type of joke. … Roman needs [a memorial], but not this kind.”
Is bringing back Black Mirror (or any other work of art that presumes to be cutting critique) really effective if the people actually engaged in an act responds to that art as well?

There's nothing new or newsworthy about this. At all.

I'd say that The Verge article is a better coverage because it actually talks about the life of the person who the bot is based on, and so actually preserves his memory.
posted by Apocryphon at 12:50 PM on April 30, 2018


I'd be more interested in the original concept in which you could feed all your text conversations etc with a person into it and it would attempt to replicate their writing style while conversing with you.
posted by Jacqueline at 12:57 PM on April 30, 2018


Ah wait, the article I quoted was on MeFi previously. Check it out to see what we thought then, a year and a half less of Black Mirrorization!
posted by Apocryphon at 1:01 PM on April 30, 2018


Is bringing back Black Mirror (or any other work of art that presumes to be cutting critique) really effective if the people actually engaged in an act responds to that art as well?

It's like asking if it matters that the entire black community has told Rachel Dolezal she isn't black, and she keeps trying to change people's minds, convinced she's in the right, and it will ALL BE OKAY if she can just finally convince every single person who has already literally told her why she is wrong that she is actually right.

The thing is, there is always a subset of the population that never learns these lessons. Always have been always will be. Literally no amount of any other human trying to teach them anything will get their head out of their ass or make them not stubborn and stupid and ignoring good advice. It's just like there will always be people who ruin public housing for everyone else. The reason the whole building has roaches/bedbugs despite 99% of the residents being meticulously clean. There always have been and always will be people who reject whole-cloth things that the vast majority if humans consider rational "good ideas." (Even fantasy stories, such as Full Metal Alchemist, operate on this knowledge that some people simply cannot accept loss and will literally wreck their lives and harm the people they love in pursuit of a literal impossibility. It's not as though we don't understand this about humans.)

However, the point stands:

Are we just going to give up critiquing everything forever because there's stubborn stupid fucks in the world?

Also, Charlie Brooker is old now, or did everyone else just forget that he had a cynical career long before Black Mirror? I guess I'm the only one who remembers Newswipe or Weekly Wipe or the Yearly Wipes or even fucking Nathan Barley (which was ahead of its bloody fucking time)? If we're going to complain about it not being "fresh" enough, are we going to ignore that his cutting criticism worked a lot better nearly 15 years ago when most of his ideas were put together? When they actually were fresh new ideas? Not the ideas of an old man who has already succeeded in the industry and is now using his clout to produce the kind of television he wants to see?

I don't usually expect the olds to have a lock on on-the-nose critique of the modern era. Usually by the time they get a voice, the ideas they had are long past being useful.
posted by deadaluspark at 1:56 PM on April 30, 2018 [2 favorites]


I've never played with one of these apps, but the article reminded me of the Judge John Hodgman episode where a woman calls out her friend who seems really into her chatbot named Tinothy. I really enjoyed that episode of the podcast, not least because the name Tinothy tickles me.
posted by Squeak Attack at 2:23 PM on April 30, 2018 [2 favorites]


Eliza or GTFO.
posted by grumpybear69 at 3:59 PM on April 30, 2018


Beyond Black Mirror there was the recent LifeAfter podcast.
posted by doctornemo at 6:18 PM on April 30, 2018


this app is Tinothy from the Judge John Hodgman podcast, just they didn’t say they brand name to avoid giving it free marketing.
posted by vogon_poet at 7:30 PM on April 30, 2018 [2 favorites]


Insert customary 'AI my arse' comment.
posted by GallonOfAlan at 10:46 PM on April 30, 2018


Literally no amount of any other human trying to teach them anything will get their head out of their ass or make them not stubborn and stupid and ignoring good advice.

Come, come, elucidate your thoughts.
posted by flabdablet at 11:43 PM on April 30, 2018


On my phones messaging app, Google will give me three options to reply to texts I get if I don't want to type my own. A lot of times it's really generic stuff, "hahahaha" options for a joke, "what you doing tonight?" during a convo about the day, kind of rote stuff. Other times, though, it seems to have a better understanding of both me and the person I'm talking to. It gets eery sometimes and I wonder if there will come a time this device will know me well enough I am able to do most text conversations just through it's auto suggestions. A chatbot AI for a living person doesn't seem like so much of a stretch if they can do it for dead onbes.
posted by GoblinHoney at 1:29 PM on May 1, 2018 [1 favorite]


« Older Die Welt als Wald   |   Only the Best! Newer »


This thread has been archived and is closed to new comments