NYT: Why Does Every Commercial for A.I. Think You’re a Moron?

Ads for consumer A.I. are struggling to imagine how the product could improve your day — unless you’re a barely functioning idiot.

I rarely see a commercial, so wonder how often people see advertising for consumer A.I..  The article doesn’t say.  Running the commercials during episodes of Murderbot or The Black Mirror might be a bad idea.

In my favorite Meta A.I. ad, a woman tells her boyfriend that she’s excited to have him meet her parents — before impishly mentioning that her father is a thermodynamicist at NASA. The befuddled boyfriend rushes home, phone in hand. “Hey Meta,” he asks. “What the heck is a thermodynamicist?”

The article criticizes this ludicrous scenario.  Other situations are presented where the A.I. saves the day.  A parent needs to buy a goldfish that resembles a pet that died and a book club host must fake an understanding of the book, Moby Dick.

The advertising agency seems to be mining mid-century TV sit-coms for wacky problems normal people face.  The agency is certainly asking an A.I. to provide situations where normal people would be helped by the assistance of an A.I..

I struggle to think of a use for A.I..  My sister uses one to generate images, like this one of Sparky as a ninja.

That is an excellent depiction of Sparky in a fantasy role, and a plausible use.  For online accounts, users might want an avatar.  In presentations, clip art or characters may be needed that don’t violate copyright laws.  As a teacher, my homework assignments and Powerpoint presentations violated copyrights about a thousand times per day.

Many people post video to social media sites.  A.I. could produce background music that isn’t going to get popped for copyright violations. 

Apparently A.I. commercials don’t address those plausible scenarios because it makes the technology seem limited.  Articles on A.I. mention that a digital assistant can help you by looking up a recipe, making airline reservations or selecting your next pair of running shoes.  None of those tasks are onerous, and I would not trust an A.I. not to just pick the third item that comes up on a Google Search.

Being retired, my uses for A.I. seem limited.  I can almost hear Jeff Bezos saying, “Choosing a Medicare Advantage option is complex and daunting.  Your digital assistant can wade through all the available plans and choose the best option for you.”

Again, I assume the A.I. just chooses the third item that comes up on a Google search. 

I can also imagine Elon Musk saying, “They are all going to fuck you.  Bypass the doctor and just get a robot physician.”

RFK Jr would add his own advice.  “Join a pickle ball team and double your intake of methylene blue, and you won’t need Medicare or a doctor.

That would prompt me to Google “methylene blue”, then wonder why food dyes are bad, but textile dyes are good.

My nephew-in-law uses A.I. to fill a blank page when he needs marketing copy.  The A.I. provides a first-cut at explaining a product or concept.  That gets him off the dime, so he can generate his own copy.  That seems plausible.  (That’s my approximation of what he explained.)

As a teacher, and probably in other careers, there is bureaucratic overhead that doesn’t help with the core objective and are probably irrelevant to everyone.  A.I. could handle those tasks. 

There are other tasks that an A.I. could have helped me with, but it feels like I would be looking for ways to use A.I. rather than doing my job better.

I’m sure there are useful tasks that an A.I. could do for me, but it seems like I’d spend more time researching how to use A.I. for that task, than it would take to complete the task.

A.I. can be used to write software, and that is something of interest to me. 

I am about 10% into home automation.  I’d like my IKEA motorized window blinds to operate on something more sophisticated than a daily schedule.  Home Assistant software would allow me to lower the blinds based on outside temperature and the position of the sun.  Only a little more hardware is required, but learning the software puts me off.

IFTTT goes beyond home automation and is a platform that allows internet-enabled devices to coordinate.  For example, a friend wrote an app to flash his LED lights red, when the Cavaliers score a basket.  That enhances his enjoyment of the game.  The programs are pretty simple.  Based on conditional statements, inputs from the internet generate outputs to the internet.  

It would be nifty to have the garage door open when the security cameras detect an expected visitor’s  car pulling into the driveway.  That shouldn’t be difficult, but I’ve never done it.

To use A.I. for programming tasks, or really any task, I would need to see someone go through the specific steps to accomplish something similar.  So far, there seem to be more examples of A.I. results, and less explanation of how those results were achieved.

There is also the privacy aspect.  Using a digital assistant like Siri, sends all of the input data to a server somewhere.  That is a concern.

I don’t know how tech companies get people to use A.I. for personal tasks.  Friends who are tech-dumb rely on their phones for all sorts of tasks.  Other friends are tech-savvy, but are comfortable with how everything works now.  I suspect that A.I. is over-hyped and advancing rapidly.  At some point, it will be mundane and advancing more slowly.  At that point, it will just be a thing we use.