What Chatbots Reveal About Our Own Shortcomings

I don't need a bot to order me a pizza or a parrot. I would love a bot that cleans the kitchen and bathroom, does my bookkeeping, laundry, and gardening, and picks up the kid from school. Get back to me in a hundred years, when you've got something really worth getting excited about.

Here comes Skynet...

Why does this article give me the creeps?

We can deplore the limitations of the new AI technologies, yet there is another truth. When the limitations are seen for what they are, intrinsic, there will not be so much concern over a one day takeover by super intelligent AI devices.
Every computer scientist understands the limitations. Stated technically the intrinsic limits are:

The NP Complete problem (likely intrinsic)
The Incompleteness Theorem ( intrinsic)
Heisenberg Uncertainty (intrinsic)

Quantum computers, if they can be developed, will not change that. Other limits are the coming end of Moore's law and the lack of understanding of the human brain. Some say it will take centuries to understand.

This leaves one kind of confused. Will we ever get the promised benefits of AI? And if not, is that bad news or good news?

The article is good summary of the visible chatbot experiements taken on by the big tech companies that all seem to have far more competent PR than R&D.

There's a lot more going on in this space than the author covered. And none of the amazing stuff is coming out of the tech titans or app-of-the-week companies.

As we used to say, back at the dawn of the computer age: Garbage In = Garbage Out. (GIGO.) Same as it ever was.

largely pointless, assistants are a dime a dozen. make a AI CEO and then I'll be impressed.

Can I say that the active GIF makes it very hard to read the story, particularly since it's popping letters at you? It gives me a headache.

A bit dystopic on the chatbots, are we, Jenna?

Most people I know, as do I, consider the phrase, " customer service" an oxymoron; how much time is already frittered away by disembodied voices that after wasting our time finally relent and let us speak with a real person, who is in a position to resolve the problem that prompted the call originally. And there's another side to this, has anyone seriously addressed the issue of what is going to happen to the workers whom all this wonderful efficiency will permanently displace?

A recent article in these Times had this quote, attributed to Google: "The A.I. resources Ms. Greene is opening up at Google are remarkable. Google’s autocomplete feature that most of us use when doing a search can instantaneously touch 500 computers in several locations as it guesses what we are looking for". Diane B. Greene is the head of Google Compute Engine. When I read this kind of drivel I cringe - an advanced superconnected computer system able to access multiple databases at high speed on the Interweb is now equated - by Google, and Facebook, and Microsoft, etc. - to intelligence. Brokerage workstations at First Boston, based on Windows 286, could do this back in the 1980s, talking to stock exchanges in Chicago, New York, Tokyo and London at the same time. ITT had a network that could connect us to the world, wasn't as fast as Vodafone, some of it used acoustic couplers made by Ozzies in Surabaia, but it all worked. So I didn't know it, but I was working on Artificial Intelligence back in the 1980s! Who knew!

I have no idea where we went off the rails, but can we at least agree that if you have to log into a provider so they can look up who you are so they can figure out what they think you want, it isn't intelligence. Let me put it this way - the day Facebook understands, when you look for shoes, you're a transgender looking for size 10.5 (UK: 9.5, EU: 44.5) heels, is the day I will tell Zuck he knows what means intelligence.

This is an effort by tech pundits spinning their own Icarus hype mixed in with FOMO phenomenon. If they get it right they will adopt the label of being visionaries and if they don't no one looses.

The spin cycle on technology hype is mirroring the dance of the dervishes. Us simple folk will be on the sidelines watching them fall on their own Damocles sword and breathe easier when real technology meets real needs to solve these problems.

For too long too many folks in Silicon Valley have created the notion that the next best things is really an artificially created & hyped unmet need and they are solving this problem. While significant tech has emerged from real world needs vast majority of tech hype is chimerical.

Wait and see... It is normal, at this stage, to have humans monitoring the service. First of all, these companies need to test whether there is a market or not, and if so invest in this technology. I'm sure in the long run bots will be 100% software. It's only the "Lean" approach that Is transforming how new products are built and launched.

As you say, all software is "only as good as its maker's imagination." And the extent of imagination in tech world is limited to the goal of turning us all into obedient & profitable over-consumers.

Seems to be a typical US scenario of what you'd want a bit to do,namely order takeaway food!
Here in Europe people cook at home or eat out occasionally.Takeaways are mainly there for for the working lunch or for people eating in the street.
Software designers in the US should realise that their culture built into AI is laughable to the rest of us.

Silicon Valley has truly run out of ambition and creativity if the biggest problem they can think to solve is "how can we make it easier to order a pizza by text when you're drunk."

I've played with bots and found them pretty useless, mostly because of Tech hubris. For example, if I could limit the speech recognition to actions I do frequently, maybe the phone would be a bit better at getting it right. Instead, if you're lucky, "Call Erik" gets a response of "There's no Eric in your phone list." Most of the time, I get a response of the sort like "Call Erik isn't playing in your area. Maybe you'd like to see Star Wars, The Force Awakens, instead."

Doing the easy stuff, like ordering a pizza from a familiar restaurant is easy. It's the more complex stuff, that I'd appreciate help with, but I really don't want to spend the time training the bot.

Just because some developers and Facebook have chosen to alter the power of the AI chatbox to that of some event captured in the breadth of a text doesn't mean these tools are broken or don't have incredible leverage...

The more we fool ourselves into thinking anything more than novelty is capable in today's young developer's mind the more we ignore those who have been working on this technology for well over a decade.

Those of us who made this possible can now only hope that all the money hasn't been wasted on the hopes of lassoing mythical unicorns and we can take some of that left over capital and produce products that perform much more meaningful tasks than ordering a pizza or scheduling an UBER ride.

Even with the so-called personal assistants like Siri, Google Now, and Cortana, it's questionable to what degree it's about actually helping the user and to what degree it's about making the user give more and more data to the corporate cloud. E.g. you can't even use Cortana unless you turn on location services which lets Microsoft track your location, even though it seems obvious that asking Cortana to set an alarm or send a text message doesn't actually require location information.

Personally, I opt out.

"Chatbots"? "Mesmerizes"? It seems to me that ALL "artificial intelligence" programs are just that: artificial......let alone intelligent. And, this whole subject and expounding of it seems to be a matter of "pulling my leg". I know. I know. This IS politics, isn't it? Better were it held and discussed during the Halloween Season. Better yet: DON'T!!!

Donald Trump has been great for comedy and I see that chatbots will be just as great for science fiction. Maybe that's how SkyNet becomes active. Someone asks his chatbot to send pizza by drone to everyone tagged on his Facebook page and a chatbot hacker sends armed drones to make pizza out of everyone with a photo on Facebook.

I'm reminded of Clippy, the little paperclip character that appeared alongside your Word document, offering help as it contextualized your intentions from what you were typing. As far as I know, most people found Clippy annoying and invasive and promptly closed his little window. Clippy eventually disappeared as the general market demonstrated how uninterested they were in virtual assistants. Seems people have short memories.

It wouldn't surprise me if many of the existing bots are Mechanical Turks in drag, as the author notes, but, if so, that should be a transitional phase. And, yes, Facebook wants its bots to let you stay on Facebook/in Messenger, but as cynical as that may seem I don't think our current an-app-for-everything approach is better. Even Facebook's bot strategy is a transitional one, as its purpose can be superseded by a better AI.

Fun stuff to watch, anyway. See: http://kimbellardblog.blogspot.com/2016/04/how-bot-that.html

In this new, and increasingly common, chatty news format, I read the entire article and still don't know what a "chatbot" is. Why do you waste our time with pointless blab? Do you think obfuscation is information?