McDonald's not lovin' its AI drive-thru experiment with IBM McDonald's is pulling out of its venture with IBM that brought AI to some of its drive-thrus. The fast food giant is slated to take the feature offline by July 26, according to Restaurant Business, which obtained an internal email sent to franchisees last week. The letter reportedly offers very little explanation for why the …
Any one who ever dealt with IBM Sales knows how this crap happens.
They get the name of a middle manager and call them up. Nice friendly chat, with a casual 'oh by the way, who do you report to?' As soon as you hang up they call back asking for your boss. They repeat their 'oh by the way, who do you report to?' spiel until they work their way up the ladder to the point where someone can authorize a multi-million dollar deal with little knowledge on the actual details.
Once the ink is dry and the people who know things hear about it, it's too late.
AI is becoming the new 'i' designation. iPad, iPhone, iAnything. It's a way to sell crap to the ignorant, which should be IBM's motto, IMHO.
If you ever listen to the way some people order, particularly the elderly it often comes across as complete nonsense and confusion. They'll ask for non-existent products, or describe them in bizarre ways and a human will somehow steer them to a valid choice. Often they'll refer to a deal they heard about that might have been for a totally different company. Countless times I've heard people order a sausage mcmuffin from a fast food chain that isn't mcdonalds, but the cahier does the conversion.
Perhaps like Siri and Alexa it was dumb as shit and could not understand spoken words properly. Less AI and better ears needed. Less ChatGPT and more Nuance/Dragon Dictate partnership needed.
Similarly I was watching a Teams meeting recording yesterday and was appalled at how bad the audio transcription was.
Last I was at McDonald’s with my elderly Dad the server at the counter was highly confused as he was asking what sort of hamburgers they had.
I think their supposition wasn't about the language processing capabilities, but instead the speech recognition capabilities. A cheap microphone in an area next to a running car and a customer talking in whatever direction they want can make for a challenging recording experience, and software that's been trained on good recordings can make mistakes even without that.
"we have captured many learnings..."
Presumably, plain English wasn't one of them.
I don't see why they need this technology anyway. Why spend shedloads of time and money trying to get a machine to do something that a kid with a headset can do moderately well on their first day in the job?
Sure, humans are moody, awkward, unreliable, want paying and are likely to quit on a whim. But, apart from the payment, so are computers, and at least the human is quick and easy to replace on the fly. Sometimes, simple is better.
McDonalds will be happy come the day there isn't a single employee in a branch, machines taking orders, machines making orders, mopping the floors and emptying the bins..
That is their goal, extracting every cent from the process for the c-suite and not a penny more than necessary for anyone else, do they care if the product drops in quality, the past decades would say that's a resounding no, do they care the level of service diminishes? Again, past experience tellls us what to expect. They believe they can market it to us enough that will overlook the worsening deal, product quality is dropping to the point it's barely describable as food and prices are going up three fold, millions are still being served.
(so, given we positively reinforce their shitty behaviour, why would they change their self interested destructive ways?)
"Why spend shedloads of time and money trying to get a machine to do something that a kid with a headset can do moderately well on their first day in the job?"
The only reason is if the machine can be cheaper, and there are times where I would prefer it. For example, the service I canceled today is something I'd have much rather done by clicking a cancel button on a website, but I had to call and talk to a person to do it (I'm guessing money was behind that one too). I imagine that successfully turning natural language into the correct restaurant order is a greater challenge than giving me a cancel button, though, so the chances of doing it with acceptable quality and still ending up cheaper are not as good.
The do say that McDonalds isn't a fast food business, its actually one of the worlds largest landlords and selling food is just a good way for their franchiser to pay their rent.
Most of the properties containing a McDonalds restaurants are owned by McDonalds Corp. and the franchisee is charged rent for the property, plus they have to buy all their stock and equipment through McDs, And they also have to pay royalties for using the McDonalnds name and branding.
So im guessing if they can persuade those 40,000+ franchisees to buy a new AI ordering system which will obviously come from corp HQ, thats a lot of profit for McDs.
That might be true, but the franchisees won't buy them unless either they're forced to or it provides them a benefit. If it ends up being more expensive than having a human doing it and they have a choice, they'll employ humans to do it. So while there's an indirection in the middle, it still mostly comes down to cost and whether they can keep a reasonable level of quality while doing it.
Yet another attempt to lessen the salary burden by this time removing a bunch of 16 year olds. I hope this AI vapourware dies a death very soon.
Governments should be pointing out the flaws in this process lest they be left with massive unemployment and surging welfare costs.
The AI pushers would have you believe we'll all have much more leisure time, but you can't afford leisure if you're unemployed.
At this rate, the only people left with jobs will be the coloured pencil department.
"Their tech doesn’t seem fake have made it to Teams yet, as the meeting recording/audio transcription is shockingly bad."
Let's hope Dragon tech never does make it into Teams, because the transcription is just great as it is. I work for a large organisation and in one Teams call, the transcription (that I preserved for posterity) shows the MD responding to one of his people's presentations with "Thank you my moron...."
SWMBO found a recipe for buns with bits of bacon in them. She offered them to her sister & husband & our daughter & family. They went down very well which is not surprising. They tasted very good and it was not at all obvious what it was that was in them.
These news articles report on two things -
1) McDonalds ending its partnership with IBM to provide this tech
2) A small number of incidents of said tech producing comical results - note that these seem to be the same handful of incidents in every article
There is implication that these two things may be linked, sure, but there is no statement from McDonalds or IBM that they are.
The fact that the trial has been running for 3 years, being expanded from 10 to 100 restraurants for the last 2 of them, but only a small number of examples exist of it going badly wrong, implies to me that it's actually been working quite well for the most part. And even if there are a handful of issues, the normal approach would be to look at those to improve the tech, not abandon it altogether.
So I suspect the reason for ending the trial is more commercial than technical, and these incidents have nothing to do with it.
And the cheery conflation of this announcement with the incidents by the news media is the usual distortion in pursuit of clicks - or as it was once called, sensationalism. Or propaganda.
The order is a couple of items from a finite list, and the customer is physically present.
So, what is easier: Having an ASR system try and understand random peoples input, against a background noise of running motors and honking horns, and THEN have a LLM or similar system figure out their intents (including corrections, repetition, the occasional sarcastic remark, etc.) and translate that into an order...
...ooooor...
...simply have an outdoor-rated tough touch-screen (like the ones used on ATMs) positioned to that it can be reached from the drivers seat through the side window?
If you want it to be extra fancy, have a Convolutional Neural Net powered visual system (see, now its also an AI system!) detect the height of the car and move the screen up and down dynamically.
You want to put your fingers all over a public touch screen and THEN eat McDonalds food with them?
Ewwwww.
I know people already do this in the restaurants but still, ewwwwww.
Use the app on your own phone at least!
If you're going to do that, have some UI testers and give them lots of time and budget for users to test with. Those screens often seem to confuse people when they're looking for some menu item that isn't there, want a specific customization option, etc. Meanwhile, if you make the most simplified interface possible, anyone who can find what they want will be delayed as they have to tap through things they aren't interested in. If they're not careful, they may find fiddling with the interface reduces the speed and their revenue.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
