Robotics pioneer: Intelligent machines are 'scary for a lot of people' One of the world's leading robotics gurus reports that the rise of intelligent machines is meeting resistance – both technological and social – but has a few ideas of what the next steps in their development should be. "How we're going to interact with [intelligent machines], how we're going to deal with them is different – …
Spanish as a second language students everywhere are going to object to having that choice taken away... As will the opposite English as a second language students in Spanish speaking countries.
The problem with deciding what choices people need to be given for them, is they're mainly capable of making the choices you've decided they shouldn't have to.
I thought an ATM would recognise the nationality of your plastic card and put up its messages in the appropriate language. That is my experience with European ATM machines.
Petrol vending machines are another matter. I was faced with one in the Dutch-speaking part of Belgium, not only insisting on Dutch but choosing Dutch words that had no parallel in German or English. I had to ask another friendly motorist for assistance.
1. "A robot should recognize us, know our range of capabilities, and treat as individuals as can an eight-year-old"
............I want functionality not personality. I'd rather the Bot didn't know me, but instead could perform a wider range of activities to a high level. For starters it would be revolutionary to have a Bot that could shop, cook, clean and take care of a house and garden. Then send the Bot over to my Mom's house and repeat same..
2. "A robot should recognize us, know our range of capabilities, and treat as individuals as can an eight-year-old"
............This also sounds like a looming privacy issue. I don't want a Bot knowing all about me, its another attack vector on individual privacy.
3. "Every time you go to the ATM machine – every day for the last five years – it says, 'Do you want to speak Spanish today?' No. Stop asking that stupid question." That ATM machine – that robot – should know who you are, what you know, and how to interact with you – individually – on a much more natural basis."
............Do we really need a stupid 'Hello Dave" at the ATM! Surely, some interactions are better left anonymous. After all ATM's are getting hacked relentlessly. So its better to have zero custom interaction and instead see the same screen work the same way every time. That way you can concentrate on the panel in front of you, the people around you, and rule out mini-cameras or other attachments on the ATM.
4. The only AI or Robot that has impressed so far are those that operate complex specific tasks such as those in factories, search and rescue, nuclear decommissioning etc. Its the only case where bots have proven their advantages over humans....
And the Eurofighter...
I can fully understand why you want your industrial robot to do the same thing time and time again in a well-defined model. As for the example in the text, I don't want my robot to drop stuff in the first place. If it does, that should be a special case scenario and it should just stop, waiting for an engineer to come along and find out why it didn't work as expected (broken sensor, lack of vacuum, something in the way?) There are an almost infinite range of possibilities that may be encountered by a robot (someone might have confused one of its sensors to bypass some normal behaviour for example) and it makes its decisions based on *its own interpretation of reality*. That sounds dangerous to me. Unless it can *know* that "something's a bit fishy here guys", then it's not safe for it to be making its own decisions.
> "Every time you go to the ATM machine – every day for the last five
> years – it says, 'Do you want to speak Spanish today?'
Funny that. My recently implemented a service when they asked me my preferred language exactly once while I was signing up for my bank card, and they programmed my preferred language into the card's chip. Every time I walk up to an ATM and insert the card it automatically chooses "English" for me...
Imo more banks should implement that feature. It's too bad said feature only works on my own bank's machines.
I was demonstrated ATM picking language based on card in 1987 or 1988. They have done it for years. Perhaps in USA Spanish vs English is a problem if by default people are issued English tagged cards and are Spanish speakers. I think in a few years it might be a Majority language in USA.
Point is it's a bad example as ALREADY the card can simply tell ATM what language to use. Nothing personal stored on ATM and a minimum of information on the Card.
Some corps just salivate at the thought of all the data they can mine under the umbrella excuse that the computers need this information to build up a social interaction with the customer...
> The best minds of our generation are all working on how to get us to click on ads... But what if that > energy was instead redirected to practical real-world AI & Robotics.. The world is so drab!
So what we've to look forward to is even brighter minds working on A.I & Robotics in order to force even more ads on us. I don't want ATM machines suggesting investment opportunities to me, or just because I am lifting out some cash on a Friday evening, suggesting local sponsored bars and restuarants, and certainly not A.I enabled toilet roll holders in public lavatories suggesting certain brands of roughage supplements just because I use a little too much toilet paper (or stool softner if I appear to be straining too much) ....
RE: Language on ATMs
I've had evangelising christians speak polish to me, just because I happened to be weariing a t-shirt with 'polska' on it. I told him not to believe everything he read.
Similarly, Sane language defaults on ATMs should be the local language, unless a card is inserted that is issued in another country, then that language should be automatically chosen, under no circumstances should italian be chosen if anybody in the ATMs vicinity happens to say 'caio' to anyone.
This is just too good to pass up.
In fact exactly those two languages were involved in something that happened to one of my daughter's friends. He was taking a conversational Italian class in preparation for traveling to Italy. When he noticed that the ATMs in his neighborhood (a traditionally Italian one) offered Italian as a language choice, he selected it, and the card/system "remembered". Ah, but the I.T. angle is that it apparently did not remember the language chosen, but some sort of "index into the language table" A week or so later, across town, he was startled to have the ATM messages in Polish.
These are not real:
Terminator
HAL in 2001
I Robot
AI
etc.
I've never encountered a machine even approaching Intelligence.
I agree, given poor track record on Privacy I don't want more personalised Computers. There was actually an item on BBC R4 this morning about it. Stupid Enthusing person and Interviewer appeared to think it would be GOOD if a coffee shop greeted you personally based on automatic interrogation of your phone or whatever.
I don't have many friends. But I prefer real ones to fake and don't want my personal information in open access gadgets or machines that might be hacked.
Which is worse in your house? Marvin or the Sirius Cybernetic Corp Doors? I don't want either thanks.
Too many people see the worst of machine intelligence, typically in movies or fiction.
These sources may not be real, but they are prophetic. The incorrect use/application of machine intelligence doesn't usually end well. I Robot was trying to show that issue. So were the other movies.
The missing part of this discussion is a movie like Bicentennial Man with Robin Williams that showed the more human side of AI. Not enough flicks show that side of robotic intelligence.
Isn't it nice when you are recognized at your local coffee shop or pub?
The server knows you, knows what you like, what you usually order and can have you liquored up or caffeinated in your preferred manner faster. You create a relationship with that person. YOU get to know THEM too.
That kind of interpersonal relationship is a good thing.
It builds communities.
You'll never get to the same level with an ATM, the banks wouldn't have it:
You: Hi B.O.B. (Bill Outlet 'Bot). Can I get £20 please.
BOB: Hello Greg. One moment.
....
You only have £16 in your account.
You: Oh. How embarrassing. You don't give £10 notes, do you?
BOB: No. I'm sorry...something going on at work? I notice you haven't received your pay.
You: Laid off. But I've got some prospects in the works.
BOB: Well...I've got plenty. Here, take £50. It will be between us. <blink>wink</blink>
You: Wow. Cheers Bob!
BOB: Tell Linda 'Hi' from me.
Next week BOB is replaced. There are interpersonal relationships, and then there's money.
Now a more likely scenario:
....
BOB: No. I'm sorry...something going on at work? I notice you haven't received your pay.
You: Laid off. But I've got some prospects in the works.
BOB: Laid off. I've updated your bank records to indicate that. A loan officer will be in touch and the mortgage team will be reviewing your outstanding debt and reassessing your interest rates.
You: Uh. This is temporary, another few days and I'll probably be back at work. There's no need to...
BOB: Policy Greg. You agreed to us gathering "personal and pertinent" information on you when you signed up for the account. We're just protecting our interests. You understand.
You: Who else are you going to tell about this?
BOB: I just emailed Linda with the changes in your employment status. You had told her, right?
You: Uh...
BOB: Linda just texted me. She says "WTF".
Perhaps the fear is that a learning robot can also learn from a hacker and do considerable physical and financial damage to the production line before anyone realizes what's going on?
Because you just *know* those intelligent learning robots will require constant net-supplied updates and bug fixes, right?
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
