‘Artificial Intelligence’ was 2016's fake news “Fake news” vexed the media classes greatly in 2016, but the tech world perfected the art long ago. With “the internet” no longer a credible vehicle for Silicon Valley’s wild fantasies and intellectual bullying of other industries – the internet clearly isn’t working for people – “AI” has taken its place. Almost everything you …
Great to see a significant outlet state that the "AI" stuff, whatever it is (and in reality very little) isn't actually AI.
Don't look behind the curtain.
Ignore the fact that the Young Lady's Primer is relying on real time human curation and a real actress.
We used to call them "Expert Systems" in 1980s. It wasn't AI then either.
"Nine years ago, the biggest financial catastrophe since the 1930s hit the world, and precisely zero bankers went to jail for it. Many kept their perks and pensions."
Technically off topic, but since a) Andrew brought it up and b) I'm in a somewhat confrontational mood after christmas with the relatives:
Well, why should there?
After all, from their (limited) perspective, they did everything right; in other words, they did exactly what it said in their job descriptions.
Logical conclusion: we're not looking at a financial crisis or a banking crisis or anything like that - we're looking at a systemic crisis. And nobody has a viable plan to get us out of it.
"Well, why should there [be anyone jailed]?
After all, from their (limited) perspective, they did everything right; in other words, they did exactly what it said in their job descriptions."
Given that some institutions were reportedly betting against their own products there would appear to be reasonable cause for someone doing jail time.
And doing what's in a job description doesn't make it legal. "I was only following orders" hasn't been an acceptable defence for the last 7 decades or so.
But IoT isn't fake news. We have had a number of exceedingly annoying DDOS attacks due to the Insecurely Developed Internet of Things (IDIoT). That's really very real, and not fake news. These things are going into all sorts of places where you don't expect them. Door bells? Children's toys? Watches with WiFi? Dog collars? Cat toys? And of course cameras and DVRs.
Not fake. Really nasty, and usually stupid.
True, but it was mostly the hype and promises how IoT is the saviour of everything that's fake I was thinking of. IoT may not be fake (just a horrible nightmare), but the promise of a glowing future if only we'd just connect everything to the internet most definitely is.
I loved the article. It reflects the sentiments that I've had myself about this for quite some time...
Parts of the advances in the use of convolutional neural networks for pattern recognition have been impressive in their own right, I think.
But, there's leap of enormous proportions from advanced pattern recognition to actual simulated intelligence. The two are on different planets altogether, so to speak, which is what most media articles seem to miss out on when they cover this topic.
It's good to finally see at least one article which puts this thing right. :)
Great article, but i think the emphasis on consumer demand is misplaced. This tech is already storming the production and logistical corner of the economy. The same could be argued for military/defense, medical, media and marketing. Commodified AI will have plenty of room to grow and mature within industry before needing to infect consumer consciousness.
... where the true masters of AI are even now using software to classify user feedback, please see Using machine learning to classify user comments on GOV.UK.
What does that involve?
They look at three features of the user feedback. That helps to classify it. With 88% accuracy in the case of one class.
And what are those three features?
Answer, "the ratio of upper case characters to total characters, the total number of characters entered in the text box, and the ratio of exclamation marks to the total number of characters".
Earlier, in Understanding more from user feedback, GDS said "this approach can also be used to tackle a range of text analysis challenges ... such as quickly understanding policy consultation responses".
No more AI scepticism, please.
The article reads as if AI is equivalent to a human capable replacement.
AI won't come with a big bang. People will realise it's here after it's happened (and maybe taken their job). The "news" is that the change is underway and accelerating. Nothing fake about that.
"AI won't come with a big bang. People will realise it's here after it's happened (and maybe taken their job)."
What people realise is that when "AI" goes into production it wasn't really anything like intelligence after all. And if it's taken the jobs of the people who were there to supply real intelligence then the business will be that much worse off in the end. What more businesses need is some RI.
You're repeating a post made 8 hours earlier so I'll repeat my response:
So it reports what they intend to save. How successful it is, including customer satisfaction, remains to be seen.
One remark in the linked article (mentioning other applications) should give them pause for thought: "incidentally, a large benefit of the software is understanding when customers get frustrated with automated systems."
yes... fortunately it's still mostly just hype, and some say the machines beating humans at various contests are often ugly kludges. I do fear the actual arrival of a capable human-level AI. And, it seems to me that the AI people (including the deep learning folks) are still somewhat clueless as to *how* consciousness should eventually emerge. No matter how much pattern-recognition power you throw at it, that alone won't result in a motivated, focused thinking machine. I would expect to see some cunning "macro architecture", possibly involving some cognitive loopy/feedback circuit, and generally macro-topologies faintly similar to a Von Neumannean CPU: search the huge memory for concepts, select one, keep it in temporary storage, search for further concepts, select some, keep them in temporary storage, search for more, run like that in cycles, with some general drives / motives / goals / aux inputs behind this basic loop (working as a filter in the selection process), with some generative/action outputs etc. Something along those lines, just possibly an order or two more complicated, multi-layered and whatnot. I actually feel that many of the basic concepts and functionality of this "sapient macro-architecture" could be tested / proven on a much smaller scale.
As MD of a company that has been in the online customer service service field for over a decade and has a Virtual Agent (chatbot, if you like) product, I agree with most of what Andrew Orlowski writes.
It has never been helpful to associate chatbots with Artificial Intelligent. They just aren't. We never do do so and it's a key point of standard customer presentation deck to differentiate the two things. Interest in chatbots has grown due to the rise of Siri, Cortana, Alexa et al and it's lazy shorthand by journalists just to slap an "AI" label on these, just because they make a token stab at communicating conversationally.
In fact, if we want to blame someone for instigating all of this, I suggest... Alan Turing. The famous Turing Test effectively invites people to draw an equivalence between AI and conversational interfaces. It's a false equivalence. The goal of machine learning should not be to mimic human beings. There are plenty of other computing problems to be solved without such red herrings. I suppose he did OK at some other things, despite that epic fail.
We're concerned that the escalating hype around AI will lead to raised expectations around chatbot interfaces that can't possibly be met.
That said, chatbots do have some benefits over more conventional interfaces. In particular, they increase engagement with a wider demographic of user by making it easy for the user to communicate their request in their own way. Everyone, from a very early age, knows how to ask a question and get an answer. We have plenty of data that suggest that users, overall, enjoy the experience.
Try thinking of chatbots as an alternative interface that everyone can use but that not everyone needs to.
Great piece. There's a tsunami of bollocks out there on this and other subjects. It's mostly rent-seekers and/or venture capital seekers filling empty airwaves with unmitigated rubbish. I have no idea why Elon Musk's every fart gets near the top of Reddit. It seems to me he's just a fantasist. No BS I could come out with ridiculous tech ideas every single day just like him. He's basically shaking down the government anyway. I don't think his ventures are actually profitable are they.
I am an AI major by the way and have understood since my undergraduate days that what people call "AI" is actually just a search problem. Expert systems are just directed search algorithms (directed by Humans selecting relevant data). Neural Networks are fuzzy pattern recognition and correlation algorithms. There's no intentionality here. No consciousness. No "thing it is like to be" an "AI" but you wouldn't know that from the media.
So yes I agree. It's all bollocks.
what people call "AI" is actually just a search problem. Expert systems are just directed search algorithms
Back in the early '80s we had library materials delivered on microfilm and an online index which took a fairly simple search language - "not" meant "don't give me anything that contains this" etc. AFAICR with a little experience the search would deliver relevant results or report nothing found.
Now we have massive online search engines - Google, Amazon etc - which can't be steered by the user. There's no effective way of eliminating crap occupying maybe the first several pages of "results", dropping inexplicably irrelevant items in the middle of the list or delivering any manner of junk rather than reporting no hits. Several decades of adding "smartness" have simply made search engines work worse.
@Robinson
I think part of the problem is what you think AI is supposed to mean. When you say of current AI that 'There's no intentionality here. No consciousness. No "thing it is like to be" an "AI" ...' it's certainly hard to disagree. So your reading of "AI" seems to equate to "human-like". But does a recognisable "intelligence" need to be human-like? Does intelligence necessarily imply consciousness (whatever that is)? Or even intentionality (whatever that is)?
I mean, I think, on the whole, that my cat is intelligent, probably has (frequently quite weird) intentions, even that she may be a thing "it is like to be" - but sometimes I think she's a dumb-ass eating and shitting automaton, programmed to exploit humans for shelter and sustenance. (Okay, I think that of some people too.) But seriously, how about a frog? A beetle? A starfish? A worm? A bacterium? (Notably, current AI doesn't come that close to the level of cognitive sophistication of any of those, even the bacterium.)
My own suspicion is that AI will creep up (is creeping up?) on us incrementally - and that it may well turn out to be quite un-human.
"Intelligence" has always been a faulty estimate of "human mental aptitude", supposedly independent of specific experience, but obviously requiring basic "normal" experience for its interrogation and measurement. When pressed to the wall, psychometricians would admit not much more than intelligence being the number which humans score on intelligence tests. With all manner of variation in definition, measurement tools, and statistical manipulations, the word and the concept it represents are so plastic that they can be made to mean just about anything. Much like "consciousness", there are about as many definitions of it as there are people writing about it, each one confident that their insight is the true insight. Pair that with human constructed, or artificial, intelligence, and one has a countably infinite number of variations ripe for exploitation by the arrogant ignoratii seeking to impress the impressionable. This is not just an amusing diversion from reality, it is a dangerous one that is absorbing more and more resources and in that process, shortchanging meaningful research and development of sorely needed advances that are really capable of benefiting humankind. This is not to say that developments in so-called AI do not have beneficial applications, only that they have little or nothing to do with intelligence as it is generally understood, much less, artificial instances of same.
Artificial Intelligence isn't
This becomes painfully obvious when my (wonderful!!) wife tries to get a request into Siri. She fills up the request with words of noise, and never gets the right answer.
Of course, there is the speech recognition software that wants to change the word "Poway" (a city in California) into two words "pow" and "way", then mangles the response when it can't find the city name.
The silly voice bots that are on the other end of the phone line are another thing. Look, we all know you have options, and yes a limited number of them, just tell us. No, I don't want to speak my credit card number when the public is listening.
And so it goes...
A lot of truth from Andrew re. the hype. But I fear the point being missed is that the mega-corps have escaped the constraint of consumer preferences. It doesn't matter to them that the AI product is crap, they'll shove it down our throats anyway. Who is happy with the Fillipino call centres - but that's what we've got ?
This is because at my workplace, we replace most of our jobs in 2005 with artificial intelligence. Imagine the AI reading the doctors handwriting with more accuracy than a human. But then it worked so well, we have not been replacing jobs since. I guess it is time for more updates, but too late for 2016....
But what is intelligence? People say that AI isn't real, it's just machine learning but what is human intelligence if not just learning? Our knowledge comes from what we have learned just like a computer. The more knowledge we give 'AI' the better it's decision making will be.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
