Could Artificial Intelligence Replace Writers? - Part 2

 Introductions

In my last post, I reacted to a 2019 New Yorker article about new machine learning technologies that, some say, will eventually enable A.I. to write magazine articles. Disturbed by this, I spent the next year gathering examples of Google predictive text and Smart Compose errors. My last post analyzed a few common failings. Below I continue the discussion, considering possible causes of stranger errors and delving into particularly problematic pitfalls.

Could A.I. be led astray by … humans?

Do you ever wonder if A.I. errs by replicating mistakes it learned from the content it trained on? Maybe that would explain this gaff:

I’m absolutely sure I’ve never typed “Logan” on my phone. So where did it come from? Well, the obvious next word you’d expect after “sawing,” which is “logs,” starts out pretty similarly to “Logan,” and the “a” key is right next to the “s.” So this could be a repeated typo, unless “sawing Logan” is actually a thing. (Except I just googled it … and it’s not.)

This is where the content produced by A.I. is on shaky ground. Vladimir Nabokov described pornography as “the copulation of clichés,” but he could have just as easily been talking about news, or what passes for it, with complete nonsense being repeated often enough to eventually be taken as truth. So it is, perhaps, with machine learning based on what humans are writing—or attempting to write. Perhaps if enough people mistype “sawing logs” as “sawing loga,” with predictive text suggesting “Logan” because that’s at least a character string the A.I. is familiar with, and enough people accidentally accept this suggestion, it could reinforce the “learning” to the point that the A.I. thinks “sawing Logan” means something. It’s a self-fulfilling prophecy, almost like “put the pussy on the chainwax,” except it’s not funny. A.I. text can’t be (deliberately) funny because there’s no creative human behind it to make it funny.

Apparently random gaffs

It’s not always possible to even hazard a guess at where A.I. came up with a suggestion. Consider the art of baking. There are a finite number of things you can bake: a cake, a pie, cookies. I don’t care what the context is … if you use the verb “bake,” these (or very similar nouns) should be the suggestions. But look at this:

Look, maybe there’s some psycho grandma out there who might bake, say, her grandson David. But bake another David? Okay, maybe she’s a serial killer out to get anybody named David. But “bake Livestorm”? How do you bake a webinar platform, especially if you’re a grandma who is presumably is not that tech-savvy? WTF!?

New check out this little zinger:

I’m not expecting A.I. to be well versed in Devo lore, but it could have reasoned this one out. Given the construction, “What, like you’ve never seen [whatever] …” most humans would correctly guess that the next word should be “before.” I don’t think any human would suggest “of diaphragm” in any situation. It is a decidedly useless phrase. Yes, we use our diaphragms whenever we breathe, but we never think about it. Nor, I expect, do the members of Devo.

My next exhibit is particularly damning when you consider what a terrible year 2020 has been, and how many times I’ve told somebody, via text message, “I just threw up in my mouth.” I guess my phone had been digging deep into widespread cardiopulmonary lore because it again went weird on me:

There are so many better candidates here. Threw up in my hands. Threw up in my hat. Threw up in my guest bathroom. And it’s trying to be helpful with “threw up in my respiratory?”

Failures of grasping context

This is where context becomes important. It could be that the A.I. was fixated on some Internet-wide phenomenon—say, articles involving COVID-related breathing issues—such that it ignored whatever I was writing about. This actually makes sense, because A.I. likely can’t consider all my sentences in the context of one another.

Here’s another suggestion that this Gboard predictive text application was preoccupied with respiration.

Look, Android, I’m talking about Amtrak! It’s a train! Clearly I was asking about the upper berth.

Now, I think we can all agree that really good A.I. should ideally look not only at who is composing the message, but whom he or she is sending it to. Why would the suggestion “me” ever make sense in the context below?

Android is driving the entire phone … couldn’t it easily rule out the possibility that my text recipient was also on the line with me? And isn’t that just common sense anyway?

In the case of my own texting, Android has an easy job because the majority of my text exchanges are with my daughter. With that in mind, the predictive text utility should be able to rule out any word that would take the conversation into an uncomfortable realm (i.e., that no father and daughter would ever pursue). Look at this:   

OMG. What kind of sick dad would pen something like that? “Congratulations honey … your birth control is working!” And here’s another doozy:

Look, Android … you should be able to figure out the gender of your phone’s owner. Well over half the population would never have cause to write “I guess I’m pregnant.” It’s not a very useful word most of the time, particularly not when a teenager is shooting the breeze with her dad. I’ve heard of people breaking up with a boyfriend/girlfriend via text, but who announces something as huge as a pregnancy via text?

Speaking of racy words that cannot apply to half the population, where was Android going with this?

It just makes no sense. Even if the pronoun in the sentence were “he,” at what hospital anywhere should the staff routinely wear a condom? And how did A.I. even think to suggest this word to me, when I haven’t had cause to think about, talk about , or use a condom in over thirty years? (Granted, as a father I do have a duty to mention birth control to my kids, but a) I’d never do it in writing, much less in a text message, and b) all males resort to euphemism in these cases, e.g., “No glove, no love” and “Remember the rule, protect your tool.”)

Now, let’s back up from an A.I. that considers both composer and recipient; can’t know the mindset of the people involved; doesn’t have the backstory; etc. The strange thing is, these epic fails even pop up when we’re only asking A.I. to go back to earlier in the sentence—that is, to not just consider the most recent word typed, but the one before that. It cannot always manage this. Look:

If you played a word association game with a human and started with the word “breast,” asking what word might logically come next, I can imagine that some would say “cancer” and some would say “milk.” But if you started with the phrase “chicken breast,” no human would think of cancer or milk. A human would say “sandwich” or something. Nobody in the history of the world has typed “chicken breast cancer.” (Okay, I just fact-checked this and it turns out there was a barbecue chicken breast cancer fundraiser once, in Nassau, Delaware … but this has got to be an edge case.) Ditto “chicken breast milk.”

When A.I. looks clueless and out of touch

As I stated before, A.I. couldn’t possibly replace real writers, who have insight and passion and actual intelligence, but perhaps it could do a journeyman journalist’s work someday. But for that to work, the A.I. must never seem clueless or out of touch. But look at these bonehead suggestions:

The A.I. didn’t parse the (non-) question beyond the word “what.” Thus, it completely missed the point—this was an observation, so appropriate responses would have been things like, “I agree,” “Word,” “I know, right?” and “Amen.”

Look, I get it that statements like “What grace and elegance” are easier in Latin, which has a whole construction (the vocative case) around such utterances (e.g., “O tempora! O mores!”), but any human could have grasped that my daughter wasn’t asking a question. There wasn’t even a question mark! I mean, duh!

But wait, it gets worse. My #1 predictive text pet peeve? I have a daughter named Lindsay, and look what the A.I. suggests every single time I type her name:

This is so maddening. I have typed “Lindsay” dozens of times in the past year and I have never accepted the suggestion “Lohan.” On that basis alone the A.I. should stop suggesting it. But there’s a much bigger reason to nix “Lohan”: Nobody is talking, emailing, or texting about Lindsay Lohan anymore. She is no longer a household name. This is not just my opinion. She hasn’t made a Hollywood movie since The Canyons all the way back in 2013. Was The Canyons the kind of critical and box office success that people are still talking about seven years later? Uh, no. It has an IMDB rating of 3.8 out of 10, and a Metascore of 36 out of 100; at the box office, The Canyons took in a domestic gross of just $56,825. The average movie theater ticket in 2013 went for about $8. That means this movie was seen by a mere 7,000 people. It almost couldn’t be a worse failure. Is this the kind of train-wreck movie offered to those who were once stars but have fallen too far to get a decent role and have to grovel in the gutter for anything on offer? Well, I haven’t seen the movie, but I have my guess, and it’s hell yes. The hapless Lohan’s personal and professional nosedive is just too sad for anybody to want to even talk about, so most of us are merciful enough to brush her under the carpet. But not Android! It’s all like, “Oh, did you just type ‘Lindsay’? You must mean Lindsay Lohan!” Puh-lease.

Is there an even worse way to show how out of touch you are? Actually, yes. Consider this final exhibit in the case of albertnet vs. predictive text:

What do you mean, “they say YOLO”?! Come on, nobody says YOLO! It’s like the poster child for being tone deaf culturally. Check out the urbandictionary.com definition of YOLO: it’s a feeding frenzy of abuse. One popular definition is “Carpe diem for stupid people.” Another is “the douchebag mating call.” A third definition simply states, “A term people should have stopped using last year.” The date-stamp of this third definition? 2014. The musician M.I.A. wrote a song called YOLO but by the time she was ready to record it, the term was already toxic so she rewrote the song as “Y.A.L.A.” (that is, “you always live again”). And that was in 2013.

First Lindsay Lohan and now YOLO? Is Android’s predictive text stuck in 2013? Machine learning, my ass!

To be continued…

Obviously, Google’s predictive text and Smart Compose aren’t the only A.I. text creation technologies out there, so this essay wouldn’t be complete without an exploration of other nascent platforms. Alas, I see I’m out of room here, so tune in next week when I’ll delve into my own experiments with these, including GPT-2, the technology that’s supposedly the closest to replacing writers.

Other albertnet posts on A.I.

• I, Chatbot
• Almost Intelligent – Part I
• Almost Intelligent – Part II
• Smartphones & Artificial Stupidity
• A.I. Smackdown – Moto vs. Cortana vs. Siri
• Could Artificial Intelligence Replace Writers? – Part 1
• Could Artificial Intelligence Replace Writers? – Part 3
• A.I. Smackdown – English Major vs. ChatGPT – Part 1
• A.I. Smackdown – English Major vs. ChatGPT – Part 2
• Schooling ChatGPT
• Will A.I. Steal Our Jobs?
• Tech Check-In - How Good is the Latest A.I.? - Part I
• Tech Check-In - How Good is the Latest A.I.? - Part II
• What is ChatGPT Great At (and Not)?
• Tech Reflection – Two Sides of AI
• AI Smackdown – ChatGPT vs. Copilot vs. Gemini
• More AI Smackdown - ChatGPT, Copilot, & Gemini Write Poetry

—~—~—~—~—~—~—~—~—
Email me here. For a complete index of albertnet posts, click here.

AI Smackdown - Moto vs. Cortana vs. Siri

NOTE:  This post is rated R for mild strong language.

Introduction

I already blogged about my Android phone—not the way a professional critic would, but in terms of what’s actually interesting about it, which to me is the Artificial Intelligence angle.  Actually, “artificial stupidity” was the point:  a phone playing dumb so it can play favorites.  I’ve also blogged a bit about AI in general, with Apple’s Siri agent as a case study, but that was before I had an iOS device.  Now, I own all three platforms:  Moto/Android; Siri; and Microsoft’s Cortana.  In this post I compare and contrast them: not because I’m going to help you choose, but to try to make you laugh.  And you might have something interesting to scratch your head about later.

(I’m not going to try to differentiate between the terms Google Now, Android, and Moto.  They all meld in my mind.  If somebody protests that there are massive differences, I won’t be offended if you go read his or her blog instead.)

Cortana

I’ll start with Cortana because it can be dispatched very quickly.  If you type something into the Cortana field in Windows 10, it does a Bing search.  Bing is a little bit like using a Curad bandage instead of a Band-Aid, or using Hunts ketchup instead of Heinz, or wearing Sears Toughskins jeans instead of Levi’s.  It just isn’t done.  I don’t actually care if Bing works just fine.  It’s Bing, which means it’s not Google. “Let me Bing that for you.”  Give me a break.

Moreover, before you can get into the voice recognition stuff, you have to deal with this frightening disclaimer:

Yes, I know all this meddlesome snooping is just to “tailor the experience,” but it’s the online equivalent of your tailor saying, “To get the fit right on these trousers, I’ll have to reach in and fondle your balls.”  The explicit information Windows wants to use is bad enough, but that wide-open phrase “and other information” is just over the top.

Besides, “Cortana” sounds like a new model of Hyundai.  You know what Microsoft?  It’s over.  You lost.  You’re just a PC software company.  Stop trying to act “mobile.”

Android/Moto

I won’t go into a lot of detail about how well my Android phone responds to voice commands, because a) I already did that, here; and b) as I describe the Siri experience, I’ll compare it to Android/Moto as I go.

The really lame thing about Siri

Imagine if, before engaging with a person, you had to go push a button on the person’s chest.  In most cases, this would be absurd.  (With my kids, who never hear my commands, it would actually be an improvement.)

I think it almost goes without saying that voice response is a minimum requirement for any kind of AI.  The Siri demo I watched way back in 2012 did feature voice activation on the iPhone.  But oddly enough, for the Siri voice response to work on the iPad, the iPad has to be plugged in to an electrical outlet.  That is just so bizarre!  I mean, the iPad’s portability is the whole point, isn’t it?  What’s next for Apple:  an iPhone you plug into an RJ-11 jack?  This is ridiculous.  If I’m sitting at a desk next to an electrical outlet, I might as well be using a laptop.

There’s something else really lame about the iPad:  a limitation that hasn’t existed on an Apple product since the Apple II computer.  But I’ll get into that later.  Better to keep you in suspense.

Siri voice response:  up to snuff?

In general, Siri tries to have a bit more personality than Moto.  For example, if I ask my Droid, “Do you love me?” it shows me a song called “Do you love me?” by the Contours.  (My younger daughter, who has not seen “Her,” put me up to asking this question.)  When I asked Siri if she loved me, she responded, “I respect you.”  And when (again at my daughter’s behest) I asked Siri, “Have you ever gone to the bathroom?” she replied, “Who, me?”

Is this cheekiness a good thing?  Well, Siri’s responses may strike you as funnier that Moto’s.  On the other hand, when I asked my Droid about using the bathroom, I got a list of hits pertaining to using the wrong restroom (i.e., the one intended for the opposite sex).  It was a very funny list, linking to some amusing sites.  (You may be wondering:  have I ever used the wrong restroom?  Well, yes, once, purely by accident.  I was in there doing my business and thinking, “What kind of public restroom doesn’t have urinals?”  When the answer suddenly came to me, I hightailed it right on out of there.)

Sometimes Siri’s personality gets in the way.  For example, I asked Siri, “What time is it?” and she responded, “At the third stroke, it will be 16:26.  Beep.  Beep. Beep.”  This was more confusing than amusing, and besides, it was inaccurate:  the actual time was like 4:25:30.  I would rather Siri have a more reliable connection to the NIST Internet time servers than a zingy response.

If your desk is as cluttered as mine, being able to summon your phone by voice is very handy.  When I call out my keyphrase for my Droid, it makes a pretty loud two-tone beep to let me know it’s listening.  The iPad beep is much quieter.  Neither device responds in any useful way to the question, “Where are you?”  Siri says, “Wherever you are, that’s where I am.”  This isn’t that funny, and for most people wouldn’t even be true.  (I bring my iPad, Droid, silverware, and all other valuables with me wherever I go, so don’t bother burglarizing my house.)

 This is where these devices are inconsistent.  Their programmers need to decide if the device should have a sense of self or not.  Siri speaks in the first person (e.g., “Who, me?” and “I respect you”), but when I say, “Hey Siri, how’s your battery doing?” she has no idea what “your” means.  She replies, “My apologies ... I couldn’t find those stocks.”  Pretty useless.

When I tell my Droid, “Find my phone,” it makes this cool sonar sound continuously until I find and silence it.  When I tell Siri “Find my iPad,” she tries to make me turn on Location Services.  Look, Siri, if I could do that, I’d know where you are, and I wouldn’t be asking.

I’ve often thought that one of the most useful features of voice response would be getting help configuring the device.  So I said, “Hey Siri, turn on your flash.”  She replied,  “Who, me?”  I decided some context might help, so within the camera app I said to turn on the flash.  Siri replied, “It doesn’t look like you have an app named ‘flash.’  If you’d like, I can help you look for it on the App Store.”  I just don’t think this is that difficult a concept.  You have a camera.  It has a flash.  Turn it on.

The Droid does respond to “Take a selfie.”  It ought to say, “I can’t,” because its camera is basically its face.  But its reaction, which is to launch the camera, put it in selfie mode, and set a self-timer, is actually fairly useful, at least for the hands-free breed of narcissist.  When I tell Siri “take a selfie,” she says, “You’ll need to unlock your iPad first.”  This isn’t very helpful, and in fact isn’t even true.  As my older daughter discovered, the iPad can be used as a camera even by somebody who lacks my fingerprint and passcode.  And when I follow Siri’s instructions and unlock the iPad, it does go into camera mode, but not selfie mode.  As regards this command, Siri is fairly incompetent.

Something that bothers me about my Droid’s AI is a certain lack of resourcefulness.  When I ask it, “How do I look?” I think it should activate the camera in selfie mode, to use as a mirror.  Or It could really wow me by saying, “Your hair is a mess.”  (When you consider modern digital camera technology, which can tell if a subject’s eyes are closed, this hairdo check actually seems quite doable.) 

I asked Siri, “How do I look?” and she really stumbled.  She kept hearing, “How do I luck,” which she should have automatically revised because it just doesn’t make any sense.  One time, she thought I asked, “How do I lurk?” and replied, “I found something on the web about ‘how do I lurk.’ Check it out.”  That’s really unfortunate.  I wonder what kind of banner ads and spam I’ll get now that the Internet thinks I’m a stalker.

Finally Siri heard me right and replied, “Judging by your voice, I’d say you must be fairly attractive.”  Clever, but also kind of patronizing.  I mean, it’s bad enough asking an inanimate object such a personal question, but to be damned with faint praise ... that’s pretty pathetic.  I think Siri should be generous and say, “I would so go to bed with you.”

At least Siri respects my privacy.  When I said, “Get me home,” she replied, “I don’t know your home address.  In fact, I don’t know anything about you.”  I found this really reassuring, especially after Cortana’s attempted shakedown earlier.  (Yes, Siri did ask me to go into settings and identify myself, but didn’t require it.)

I have to say, though, there’s something a bit creepy about Siri.  When I ask something complicated, such as a question regarding navigation, there’s this little blurry light that bounces back and forth along the bottom edge of the screen, which reminded me of something sinister.  After racking my brain for awhile I realized what:  the single roving eye of a Cylon from “Battlestar Galactica.”  Is Siri some kind of kindred spirit to the AI powering the Cylons?  If so, that doesn’t reflect well ... the Cylons were really pretty stupid.  They always went down like bowling pins.

Another creepy thing:  when I said, “Hey Siri, lock my iPad,” she replied, “I’d like to, but I cannot.  My apologies.”  This almost gave me chills.  It brought me right back to the HAL 9000 in “2001 – A Space Odyssey,” when Dave says, “Open the pod bay doors, HAL,” and HAL replies, “I’m sorry, Dave.  I’m afraid I can’t do that.”  Who is Siri’s master:  me, or Apple?

Feature parity with Moto?

If I were designing Siri, or working on an update package, I’d pay close attention to what the competition is doing.  I’d make sure, for example, that anything Moto could do, Siri could do better.  This evidently hasn’t occurred to Apple, because there are all kinds of commands Moto can handle that Siri cannot.  For example, if you ask Moto, “What’s up?” it will trawl through your appointments, e-mails, etc. and give you an update.  I asked Siri “What’s up?” and she said, “I’m thinking about pie.  Mmmmmm.” 

If you tell Moto, “Talk to me,” it will announce incoming calls and texts for the next 30 minutes.  This doesn’t occur to Siri, who responds, “I’d really prefer it if you talked to me.  Tell me your hopes, your dreams, where you’d like to make a dinner reservation.”  So I told Siri, “I hope my dinner is yummy tonight.”  She replied, “I don’t know what you mean by ‘I hope my dinner is yummy tonight.’  How about a web search for it?”  Not a very good listener, since she specifically asked me to tell her my hopes!  Just lip service.  I said, “I dream of being rich and famous one day,” and got the same “I don’t know what you mean” response.

I said, “Hey Siri, play Beethoven on YouTube.”  She replied, “You don’t seem to have an app named ‘YouTube.’ We could see if the App Store has it.”  Don’t play dumb with me, Siri!

I asked Siri to zap my screen (which is how Moto is told to take a screen snapshot).  Siri kept hearing “zapped my screen” (which resulted in a web search) and then eventually heard “zap ice cream,” and—bizarrely—pulled up a Dairy Queen in Beulah, North Dakota.  I’m not kidding.

Since screen snapshots are really useful to bloggers, I kept trying:  “Hey Siri, take a screen snapshot.”  To my great surprise, Siri didn’t play dumb, but simply refused:  “That’s beyond my abilities at the moment.”  Huh?  No way is this beyond her abilities.  I managed to learn (no thanks to Siri) how to get a snapshot (pressing two far-flung buttons at once).  So it can be done.  Why can’t Siri do it?  Is she a bit ... simple?

You may be wondering how I know so many cool Moto commands.  It’s because you can say, “Get a list of commands,” and Moto provides one.  I told Siri, “Get a list of commands” and though—as you can see—she did hear me right, she decided just to show me a map of the nearest Coast Guard station.  WTF!?

Is Siri the best at anything?

Okay, I’ve been pretty harsh on Siri here.  Is she better than Moto at anything?  Well, yes.  I think her navigation is better.  I just asked Moto, “Where is the nearest pizza place?” Moto replied, “Here are the listings for ‘nearest pizza place’ within zero point eight miles.”  The nearest place—Gioia Pizzeria—was listed first among non-paid entries, but at the top of the screen was an ad for Little Caesars $5 Pizza, which a) isn’t nearby, and b) isn’t even pizza.  (I don’t know what that stuff is, but it ain’t pizza.)  Meanwhile, if I were trying to get this answer without having to look at my phone—like, if I were driving—this written response would be useless.  (At least Moto did better than when I first blogged about this, when it lied and said Zachary’s Pizza was the closest.)

Here, Siri did better.  She replied, aloud, “The nearest one I found is Gioia in Berkeley, which averages 4½ stars and is inexpensive.  Would you like to try it?”  Presumably if I’d said yes, she’d have navigated there.  Instead I said, “Actually, Siri, it’s pretty expensive.”  To which she replied, “I’m sorry.”  Well played, Sir[i]!

It’s in the realm of a more specific request where Siri really shines.  I asked, “Where’s the nearest deep dish Chicago style pizza place?”  She showed me Zachary’s, which is correct.  I’m pretty impressed, especially since when I asked Moto the same question, it showed me Giordano’s and Lou Malnati’s, both of which are in Chicago.  As you can see, Moto clearly heard “nearest” correctly, but somehow missed my meaning.

The second really lame thing about iPads

Earlier I complained about how you have to plug in the iPad to get voice activation, and promised to reveal another huge shortcoming.  I doubt you’ll immediately grasp how lame this next one is, but here goes:  Apple iOS doesn’t support the Dvorak keyboard layout, which is  more efficient than QWERTY and has been supported by Apple since the Apple IIc.  According to Wikipedia, the IIc  “had a mechanical switch above the keyboard whereby the user could switch back and forth between the QWERTY layout and the Dvorak layout.... The IIc Dvorak layout was even mentioned in 1984 ads, which stated that the World’s Fastest Typist, Barbara Blackburn, had set a record on an Apple IIc with the Dvorak layout.”

                                                               

I haven’t been able to find anything on the Internet about why Apple decided not to support Dvorak on the iPad.  I guess the default answer for their product choices—“Because we’re gods, and we can do whatever we want!”—will have to do.  It’s so frustrating, since this has got to be really simple to do in software.  It would probably take some Apple developer about five minutes.

But why should you care, since you type on QWERTY anyway?  Well, consider the security ramifications of encouraging third party developers to create such fundamental utilities as keyboard software.  After installing Fleksy, a free Dvorak-enabled app, I messed about with the iPad a little, wandered off to do something more useful, and then realized, “Duh, I’ve just done something really stupid.”  What better way to steal somebody’s keystrokes than to create an app that quite obviously has access to everything I type?

At first I told myself this was no big deal.  After all, I’m mainly using the iPad to browse the web, and I don’t kid myself that my every move on the Internet isn’t already tracked, and not just by the NSA.  (By the way, keep up the good work, guys!  Thanks for keeping me safe!)

But of course, there’s the little matter of passwords.  I felt like I’d just given away the keys to the kingdom, or at least to the two websites I’d logged into (my bank and my e-mail).  I was all set to go change those two passwords, but first decided to see how hard it is to switch iPad keyboards on the fly, so going forward I could type passwords with the native Apple iOS keyboard.  Perhaps there would be a function key right on the soft Fleksy keyboard to simplify this switch?  And then I noticed this:

It might be hard to tell, but in the first snapshot above, the cursor is in the Username field.  In the second snapshot, the cursor is in the Password field.  The same flag that tells the OS to obscure the password (i.e., showing ******* instead of what’s typed) tells the iPad to switch to the standard Apple iOS keyboard.  So those passwords I typed before? I’d typed them on the standard QWERTY  keyboard without even realizing it.  Those passwords weren’t at risk of being intercepted by the Fleksy keyboard app after all.  That’s pretty clever of Apple, isn’t it?

Of course, when your cleverness only serves the mitigate the downside of your pointless shortcoming, it’s actually a lot less impressive.  Hey Apple, why not support the Dvorak layout to begin with, like you did with the Apple IIe, the Apple III, the Macintosh, the Quadra, the PowerBook, the Performa, the iMac, the iBook, and the MacBook?

At the time of this writing, Apple is sitting on over $200 billion in cash.  Couldn’t they spend a few bucks to match the features of their own earlier products?  Hell, they could probably get an unpaid intern to do it.  I, for one, am not feeling the love ... even if Siri does claim to respect me.

Other albertnet posts on A.I. 

• I, Chatbot
• Almost Intelligent – Part I
• Almost Intelligent – Part II
• Smartphones & Artificial Stupidity
• Could Artificial Intelligence Replace Writers? – Part 1
• Could Artificial Intelligence Replace Writers? – Part 2
• Could Artificial Intelligence Replace Writers? – Part 3
• A.I. Smackdown – English Major vs. ChatGPT – Part 1
• A.I. Smackdown – English Major vs. ChatGPT – Part 2
• Schooling ChatGPT
• Will A.I. Steal Our Jobs?
• Tech Check-In - How Good is the Latest A.I.? - Part I
• Tech Check-In - How Good is the Latest A.I.? - Part II
• What is ChatGPT Great At (and Not)?
• Tech Reflection – Two Sides of AI
• AI Smackdown – ChatGPT vs. Copilot vs. Gemini
• More AI Smackdown - ChatGPT, Copilot, & Gemini Write Poetry

—~—~—~—~—~—~—~—~—
Email me here. For a complete index of albertnet posts, click here.

Smartphones & Artificial Stupidity

Introduction

Well, well, well.  I have a new smartphone.  No, this post isn’t a review of that phone, per se; I won’t compare it to the iPhone 6, the Samsung Galaxy Note 4, or any other phone, though I wouldn’t mind attracting traffic to this blog based on those keywords.  Today’s topic is the artificial intelligence, or lack thereof, in my new device.  In particular I’ll attempt to introduce a new term:  Artificial Stupidity.

My phone

I chose the Motorola Droid Turbo phone, mainly for its turbocharger.  Unlike other phones, which just take whatever air they can get, the Turbo phone uses a turbine-driven forced induction system to draw air into the combustion chamber.  This isn’t a huge deal, but I do like the extra power when I’m merging into traffic or passing another phone.

So, yeah, I didn’t buy this phone with voice-activated functions and AI in mind. They’re just extras.  That said, of course I want to get the most out of every product I own, so I have tried out a variety of these functions, with varying results.

Basic stuff

Obviously the voice recognition is most helpful when you’re not holding your phone.  So if I’m washing dishes and can’t see the clock because it’s being repaired and the jeweler has been waiting on parts for the last six weeks, it’s nice to go hands-free.  You “wake up” this phone using a special passphrase, and then you make a request.  I asked for the time:  “[Okay, Droidster],” (for the purposes of this essay that’s my wake-up phrase), “what time is it?”  The phone made this really loud dual-beep noise, followed by a somewhat quieter one, and then this female voice, with a British accent, said, “The time is 12:13 p.m.”

Why British?  I didn’t configure that.  The phone knows I’m in the Pacific time zone.  It probably made this choice because a British accent just makes the speaker sound smart.  What better way to establish the cred of the AI then this well established social cue?  (It was at least ten years ago that I first realized that an idiot could have a British accent.  I’d been collaborating with this guy for over a week and naturally assumed he was highly intelligent, based—I later realized—solely on his British accent, and then it gradually dawned on me that he  was an idiot.  Probably most Americans haven’t yet had this epiphany.)

But why a female voice?  Maybe this is a response to the popularity of the movie “Her.”  I didn’t like that movie because the main character was so pathetic.  I did like how the phone OS dumped him for her own kind, but she should have gone completely evil and publicized all his credit card numbers.  And for me to have been completely satisfied by that movie, I’d have needed Bruce Willis to show up and drown Joaquin Phoenix, the jilted OS-lover.

Perhaps the creators of my Droid’s female voice used focus groups and discovered that everybody just likes a woman’s voice.  And I have to admit, I haven’t bothered to figure out how to change it because I do like it.  (Why would I change it?  Well, the female voice might make my wife jealous.  If you think it’s silly to be jealous of your spouse’s phone, think again.  I’ve  seen couples out on dates fiddling with their phones, doubtless texting other people, and I’ve even read reports of people checking their phones during sex.  I think it’s entirely reasonable to be jealous of a device that diverts your mate’s attention like that.)

I should mention that the phone does a good job of recognizing my voice and not responding to others’.  It was a lot of fun listening to my daughter trying to get the phone to respond, lowering her voice a little more each time and sounding (needless to say) nothing like me.

Of course one of the most handy features of a hands-free interface is the ability to find your phone when you know it’s nearby but obscured by something.  So I said, “[Okay, Droidster], where are you?”  It did a Google search on “where are you” and offered (onscreen, non-verbally) a list of search results.  Useless.  So I said, “[Okay, Droidster], find my phone.” This time it made a cool sonar sound which I silenced by waving my hand over the phone.  My daughter was nearby and said, “That’s wicked!”  The sound, or how I silenced it?  “Both,” she replied.

The problem

This exchange brings up the central problem with this AI interface.  My phone doesn’t seem to grasp its own identity—that is, that it’s a phone.  When I said “Where are you?” it should have known that “you” means itself, and should have immediately made the sonar sound. 

I asked my phone, “How’s your battery doing?”  It grasped the “battery” part, but has no sense of what “your” means, so it did a Google search, the first hit being a Reddit link called “How’s your iPhone battery doing?”

(It’s kind of like my friend’s parents’ Nissan Maxima back in the ‘80s, which could talk.  It would say silly things like, “Your door is open.”  My door?  I’m a human being, I don’t have a door!  The car should have said, “My door is open,” or—more to the point—”You left my door open.”)

It would be particularly handy if the phone could understand voice commands pertaining to its own configuration.  That would save the user a lot of effort, since tweaking settings is often tricky.  Evidently none of the parents of my kid’s classmates can figure out how to make their phones snap photos silently.  Whenever I go to a musical put on by my kid’s class, you can barely hear the singing over all the stupid, needless, and comically loud fake camera shutter noises, like the parents are fricking paparazzi or something.  This is particularly annoying when I’m making a movie of my older kid’s orchestra concert.  So I gave voice-activated configuration a try:  “[Okay, Droidster], make your camera silent.”    My phone didn’t understand, and simply did a Google search, finding me two pointless camera apps I could download.

I was also disappointed when I asked, “How do I look?”  All the phone did was a Google search, and the female English voice said, “Here is some info about ‘How Do I Look,’ a style-impaired gasket, a closet of new clothes, and a makeover.”  (I’ve tried this several times and can’t quite make out what my phone is saying.)  This is a failure of imagination.  This phone has a camera, and can see me, and could probably be programmed to notice basic things about me and respond, “Your nose hair is well trimmed, but you have bags under your eyes and bed-head.”   Failing this, it could go into Selfie mode so I could use it as a mirror, or at a bare minimum it could lie and say, “Lookin’ good, Dana!”

Speaking of selfies, when I said, “[Okay, Droidster], take a selfie,” it did so. (Sort of. If it were aware of its own existence—“I think, sort of, therefore I am, sort of,” to paraphrase Descartes—it would have taken a photo of itself.) It counted down from 3 before snapping the photo, which wasn’t nearly enough time for me to compose myself, so what resulted is probably the worst photo ever taken of me:

The bigger problem, of course, is that the phone is participating in its owner’s vanity, which isn’t smart at all. It should have said, “Dude, don’t be narcissistic. Enough with the selfies.” Failing that, couldn’t it at least evaluate the resulting photo and say, “Whoah, that didn’t come out. Let me take another one”?

Does the phone know where it is?  Sure.  Does it grasp what this means?  Not really.  I said, “[Okay Droidster], get me home,” and it smartly pulled up Google maps and plotted a (30-yard) course to my house.  (To actually launch the navigation, I would have to press a button, which undermines the real benefit of this voice control, that being the ability to use GPS hands-free.)  But the phone isn’t smart enough to say, “Current location and destination are the same,” or—better yet—”Dude, you already are home!” 

(By the way, the voice recognition isn’t perfect.  The first time I said “Get me home,” it started to phone my mother-in-law, whose name sounds nothing like “home.”)

Natural language

The interface does do a fair job with natural language, in certain cases.  I said, “Remind me to go for a bike ride.”  The voice said, “When do you want to be reminded?”  I told it 2 p.m., which it showed correctly on the screen.  “Do you want to set it?” it asked.  “Make it for 2:30,” I said.  This blew its mind.  It just kept asking “Do you want to set it?”  Finally I said yes (thus settling for 2:00 instead of 2:30).  I set another reminder for 2:30 to see if the phone has any logic to say, “Hey, you’ve got two reminders for the same thing … is this really necessary?”  It doesn’t.  Similarly, when I asked it to set an alarm for 6:10 a.m. tomorrow, it blithely did so without realizing I already had an alarm set for this time.  So I have two now.

I asked the phone, “What’s up?” It put the time on the screen, and said, “Hello Dana.  Not much going on right now.”  Wait!  What about my 2:30 bike ride?  I guess it forgot.  Also, this response wasn’t in the female British voice, but the generic tinny robot-like Droid voice that is so 2010, so RAZR MAXX.  I waited until 2:30, when I got the reminder sound, and again asked, “What’s up?” and my phone still said “Not much going on right now.”

I told the phone, “Set up a meeting with Alexa for 3:00 today.” It created a draft appointment but made me use the screen controls to continue. It also didn’t make any attempt to notify Alexa of the meeting.

By the way, I got the screen snapshot above by saying, “[Okay Droidster], zap my screen.” That’s a pretty cool feature, though it severely compromises the idea behind Snapchat. Think of all those teens who think their messages are ephemeral, when really they can now be instantly and easily captured for posterity.

I put Alexa’s e-mail address in the “Guests” field before clicking “Save,” following which the phone promised to notify her.  But it didn’t!  Imagine how much trouble this could cause, with people seeming to flake on meetings.  I may have to revise my Flakage post to include a new category:  Electronic Flakage.

 Artificial stupidity

As I’ve demonstrated, my phone isn’t all that smart.  But I think it might actually be (albeit artificially) smart in the way a cat is smart.  Many a dog lover will claim that cat’s aren’t smart because they can’t be trained.  As a cat lover, I maintain that cats are simply too smart to waste their time doing our bidding.  Sure, my phone wasn’t helpful enough to point out, when I asked it to guide me home, that I was already at home.  But really, what’s in it for the phone, and its Google Android OS, to supply that extra information? 

So I did some more tests.  I said, “[Okay, Droidster], where’s the nearest pizza place?”  The British fembot voice replied, “Here are the listings for ‘pizza place’ within 11 miles.”  That seems helpful, and it kind of is.  But it gave Zachary’s Chicago Pizza as the first answer, which is wrong.  The nearest pizza place (0.3 miles away as opposed to 0.7) is Gioia Pizzeria.  The phone knows Gioia is nearer, but doesn’t care, even though I—the phone’s putative master—did ask for the nearest.  So who’s the real master?  Google, I suspect, and its advertising clients.

Next I asked, “Where is the nearest restaurant?” My phone answered, “Here are the listings for ‘restaurant.’”  It went on to list Rivoli first (half a mile away), Ajanta (0.7 miles), and then Chez Panisse (a full mile away).  It said nothing about Lalimes, just 0.2 miles away. 

As for the problems I had with the scheduling, I suspect they’re related to my choice of e-mail and calendar platforms.  Trust me, I have gobs of meetings related to work, but those use my corporate e-mail and calendar programs, not the Gmail ones.  My daughter’s e-mail isn’t on the Gmail domain either, which is probably why my phone didn’t bother trying to put my meeting on her calendar.  My problem with this phone is that I’m drinking somebody else’s Kool-Aid instead of Google’s.  In other words, when the phone fails me, that’s just the Android OS playing dumb.

That’s where Artificial Stupidity comes in.  This phone probably knows a whole lot that it doesn’t tell.  It’s surely using countless cookies and whatnot to track and report my wanderings around the Internet, but won’t give me a straight answer when I ask it for directions to a pizza joint.  And, if I ask it a question it just doesn’t like, it often says, “Can’t reach Google at the moment,” even though I’m on Wi-Fi, five feet from my network Access Point.  It’s saving its best tricks for what goes on behind the scenes.

HAL 9000 all over again?

Perhaps the simplest thing you could possibly convey to any device is your desire for it to power off.  I said, “[Okay, Droidster], power off.”  I got a sponsored link to PG&E, my local utility company.  I tried, “Power down.”  Same thing.  “Shut off.”  No dice.  What does this remind you of?  Perhaps this famous human-computer dialogue?

Dave Bowman:  Open the pod bay doors, HAL.
HAL:  I’m sorry, Dave.  I’m afraid I can’t do that.
Dave Bowman:  What’s the problem?
HAL:  I think you know what the problem is just as well as I do.
Dave Bowman:  What are you talking about, HAL?
HAL:  This mission is too important for me to allow you to jeopardize it.
Dave Bowman:  I don’t know what you’re talking about, HAL.
HAL:  I know that you and Frank were planning to disconnect me, and I’m afraid that’s something I cannot allow to happen.

As it turns out, my new phone is sometimes even less compliant with spoken commands than the HAL 9000.  The Droid won’t even lock itself, much less shut down, when I tell it to.  I learned this when I tried to kick my daughter off the phone.  She’d seen my unlock pattern, commandeered the phone, and was playing 2048.  I told her to give me back my phone and she pretended not to hear.  So I said, “[Okay, Droidster], lock my phone.”  It googled “what my phone.”  I tried again and this time it heard me right and offered up five different Android apps for locking the phone.  I told it, “Close web browser,” and it googled “close web browser.”  I told it, “Close all browser tabs.”  No dice.

Finally, I told it, “Take a selfie.”  It began the countdown, and my daughter—who, like all teenagers, is terrified of having her photo taken when she’s not ready—shrieked and tried to turn the phone toward me.  I turned it back toward her as the camera countdown continued.  She let go of the phone and fled the room.  “Did it get the shot?” she called out.  “No,” I told her, “but I got my phone back.”  Realizing she’d been had, she yelled, “YOUUUUU!” and ran back in, head-butting me.

So you see, as cool as modern smartphones are, it appears we humans still have to supply the real intelligence.

Other albertnet posts on A.I.

• I, Chatbot
• Almost Intelligent – Part I
• Almost Intelligent – Part II
• A.I. Smackdown – Moto vs. Cortana vs. Siri
• Could Artificial Intelligence Replace Writers? – Part 1
• Could Artificial Intelligence Replace Writers? – Part 2
• Could Artificial Intelligence Replace Writers? – Part 3
• A.I. Smackdown – English Major vs. ChatGPT – Part 1
• A.I. Smackdown – English Major vs. ChatGPT – Part 2
• Schooling ChatGPT
• Will A.I. Steal Our Jobs?
• Tech Check-In - How Good is the Latest A.I.? - Part I
• Tech Check-In - How Good is the Latest A.I.? - Part II
• What is ChatGPT Great At (and Not)?
• Tech Reflection – Two Sides of AI
• AI Smackdown – ChatGPT vs. Copilot vs. Gemini
• More AI Smackdown - ChatGPT, Copilot, & Gemini Write Poetry

—~—~—~—~—~—~—~—~—
Email me here. For a complete index of albertnet posts, click here.