RSS

Category Archives: artificial intelligence

Is Siri artificially intelligent or just a robot?

Having just returned from a trip to New York on Sunday evening, I haven’t had much time to play with the week’s biggest and hottest new gadget - the iPhone 4S - but I have been able to formulate some initial impressions, especially in regards to its main new feature: the Siri personal assistant.

First, the basics. Yup, the iPhone 4S works as advertised. It’s faster, slicker and generally better than its predecessor, the iPhone 4. Some nifty additions to the operating system make things easier, like you can fire the thing up initially without having to connect it to your computer and you can share iTunes purchases between devices by turning on the iCloud storage option. Both options do a lot for eliminating cables and computers from the iPhone equation.

I particularly like the camera as well. The iPhone 4 had the best camera on any phone I’d tried so far and the 4S is yet another step up. Apple is continuing to strengthen the case for leaving the full camera behind and simply relying on a phone to take photos, at least in casual situations.

Much of the brouhaha over the new device, however, lies with Siri, the voice-recognition feature that can tell the user about everything from the weather to sports scores to scheduled meetings. In Canada, Siri’s capabilities are somewhat more limited for now - it can’t, for example, give directions because it is not yet programmed with Canadian maps.

I have to admit to being fairly skeptical when Apple CEO Tim Cook unveiled the feature a few weeks ago. It seemed similar to what’s already available on Android phones, which allow voice searches and other functions, while voice recognition in general is a technology that has come a long way but is still plagued by problems.

Siri is, however, remarkably accurate. Despite never having been exposed to my voice before, the tool got all sorts of proper names (mostly) correct, even if it didn’t always know the correct answer. For instance, I asked “What is Neil Gaiman’s best book?” Siri didn’t know and suggested a web search (the right answer, of course, is American Gods), but it did at least identify the author’s name correctly, albeit spelled with a “y” instead of an “i.”

The tool works well with features embedded in the iPhone, such as weather forecasts, calling up contacts, scheduling notifications and launching media. Indeed, after a few hours of playing with it, I may just be ready to say it’s more accurate than anything I’ve seen on Android. And that’s impressive, given how much effort Google has been exerting to gather voice samples.

What really got my attention over the weekend, however, was all the fun and crazy stuff people were trying with Siri, which is now being documented by various websites including “Shit That Siri Says.” Some gems from that site include people asking about - and getting funny answers on - topics ranging from boobs, glory holes and Spongebob Squarepants.

That got me wondering as to the level of Siri’s artificial intelligence. It’s a loaded term that means different things to different people, so I asked Siri to define it. Here’s what I got back:

Under that definition, which is as good as any, the key to AI is the ability to creatively solve a problem. There’s no denying that Siri’s ability to recognize and translate voice plus grammar into usable data or actions qualifies. In that sense, Siri possesses what seems to be a good level of artificial intelligence.

However, with the sort of stuff showing up on the websites above, a good portion of Siri’s capabilities are likely simple programmed responses. It’s doubtful that even IBM’s Watson supercomputer, which not too long ago whupped human butt on Jeopardy, could construct such creative and funny responses as, “No comment, douche bag” to questions such as “Are you menstruating?”

In such regards, Siri is more of a programmed robot than a thinking entity. Somebody somewhere - or more likely, many people somewhere - have spent a good deal of time anticipating and then programming Siri with potential questions and their respective answers, humourous or otherwise.

None of that is meant to take away from the tool. It’s a cool little bell/whistle to have on a phone and it should be fascinating to see if it raises the stakes in this particular aspect of the smartphone war with Google.

Incidentally, where did Siri come from? The military, of course. As the video here explains, Apple’s tool started as a DARPA project called Cognitive Assistant that Learns and Organizes (CALO). If you’ve read Sex, Bombs and Burgers, that shouldn’t come as a surprise.

 
1 Comment

Posted by on October 17, 2011 in apple, artificial intelligence, iphone

 

2021: Computers surpass human brains

I promised in January that I’d write up my technology predictions for 2011, but given that we’re already a quarter done the year, that would seem kind of foolish. I was actually always leaning toward taking a further-out view into the future anyway, as it’s not only more fun but also harder to be wrong (ahem). In that vein, given that I’m currently on vacation in Thailand, I thought I’d kick off 10 days of predictions of where we’ll be 10 years from now. I hope you enjoy.

Let’s start with computers, which are a hot topic right now given the recent performance of IBM’s Watson on Jeopardy. Watson’s ability to make its super-smart human opponents look like tools on the game show has been thoroughly discussed and analyzed. One of the more interesting takes appeared in the Ottawa Citizen, which postulated that Watson’s performance was not completely unlike the Wizard of Oz - it was actually an illusion that required someone behind the curtain to pull the strings.

Most commentators, however, gravitated toward how Watson figures in to the issue of the Singularity, or the upcoming point in time when computer intelligence outstrips our own, thereby resulting in a new world that is currently hard for our mere human brains to imagine. The Singularity is something a number of futurists, most notably Ray Kurzweil, have pondered for years. I’ve been knee deep in it for several months, since I read Kurzweil’s latest book, The Singularity is Near, and it will actually be somewhat central to my next book (check out my interview with Kurzweil from a few months back).

This event won’t arrive in the next 10 years - Kurzweil thinks 2045 is the year when everything will change - but what is very likely to happen is that computers will in fact become more powerful than humans. And by powerful, I’m talking strictly about their computational ability. Super-computers such as Watson, which is capable of 80 trillion operations per second, will continue to see the benefits of Moore’s Law - where computational ability roughly doubles every 18 months - for the next decade. By that measure, super-computers in 2021 will be capable of more than five quadrillion calculations per second.

Many very, very smart people have tried to estimate the computational ability of the human brain based on what we know about the organ - how many neurons it has, how fast they operate, and so on - and there have been various conclusions. The one I’ve seen most often seems to estimate the human brain as capable of 100 million MIPS, or millions of instructions per second. I’m no mathematician, but that’s less than what the most powerful computers will be capable of in 10 years.

What does this mean? Well, there are a few things to keep in mind. Firstly, much of our brain’s computational ability is spent on running the body - regulating our organs, making sure we blink, walking and chewing gum at the same time, etc. One idea that fascinates me is what might happen if we could perhaps outsource some of these mundane tasks to a computer that we wear? If we freed up most of our brain power, could we possibly orient it toward more productive uses? Perhaps, but that’s getting outside the scope of the next 10 years.

Indeed, although super-computers that have more brute power than the brain are likely in the next decade, their true impacts won’t be felt just yet because of their size and cost. Watson, for example, is actually made up of 10 refrigerator sized stacks and cost several billion dollars. It’ll likely take another decade just to get such computers down in size and cost to the point where they can have a mainstream effect. Of course, by 2021, we’ll have Watsons on our desktops and smartphones, so it’s all relative.

These super-computers are already affecting change in the world, although it’s in ways we don’t really perceive. If you don’t believe me, open up a new browser tab and do a Google search on something, anything. If you’re amazed that the search engine seems to complete your query before you even type it, well, welcome to super-computers in action. Faster and smarter machines are already analyzing and interpreting our data for us. This sort of stuff is going to be all around us over the next decade.

The real fun begins when these super-computers start to interact with human brains. Our organic matter doesn’t directly benefit from Moore’s Law like computers do, but as I’ll detail in tomorrow’s prediction, the brain will in fact be redefined over the next 10 years.

 
1 Comment

Posted by on February 28, 2011 in artificial intelligence, ray kurzweil

 

Jeopardy’s Watson computer still needs humans

If you didn’t tune in to Jeopardy last night to see IBM’s “Watson” supercomputer take on a pair of the show’s past human uber-champions, you have another chance tonight in part two of what is history in the making.

The first half of last night’s episode was incredible to watch - the computer literally kicked the crap out of human competitors Ken Jennings (who once had a 74-game winning streak) and Brad Rutter during the first half of the show, only to come back down to earth during the second half. After breezing through the easy questions, Watson actually fielded a number of wrong answers - gasp! Here’s a promo video explaining what it’s all about:

The cynic might think this whole computer-gimmick-on-Jeopardy thing is just a big commercial for IBM, and to some extent that’s correct. But there really is something far more important going on here. Just as Art Linkletter used the massive UNIVAC computer on his show in the 1950s to pair up men and women, so too is Watson intended to demonstrate just what is possible with today’s computing power. As the IBM engineers say in the video, Watson can ultimately be applied to all sorts of real-world applications, such as diagnosing patients. Indeed, the military is working on a similar program called the Personalized Assistant that Learns (PAL), which can interpret data and suggest courses of action to its human operators.

Watson and PAL are the products of Moore’s Law, where the processing power of computers doubles roughly every 18 months - a phenomenon that has essentially been true since the invention of the transistor back in the 1940s. As the Watson/Jeopardy hype built over the weekend, many people joked about the coming robot apocalypse - a time long foreseen in science-fiction when the computers become smarter than the humans and eventually take over the world.

The world domination part is unlikely to happen for a host of reasons, but computers surpassing us in intelligence is inevitable. Moore’s Law, or what futurist Ray Kurzweil calls the law of accelerating returns, essentially guarantees it.

Coincidentally, I happened to recently speak with Jim Osborn, the executive director of the Quality of Life Technology Center at Pittsburgh’s Carnegie Mellon University, a hotspot of robotics and artificial intelligence research. We chatted about a story on robots that I’ve been working on and, of course, the topic of robotic AI came up. He said that thanks to Moore’s Law and the law of accelerating returns, robotics problems - things like getting a robot to walk properly or grip an item correctly - are being solved with amazing frequency. However, despite that, the role of humans in determining the intelligence level of a robot is still extremely important because we’re the ones the give it the tools it needs to learn.

“It’s not just a matter of waiting for a faster computer to do the same algorithm just that much quicker, we still need fundamental improvements in the algorithms,” he said. “You don’t want to run a crummy algorithm on a better computer.”

In any event, if you want to see something amazing, tune in to see Watson in action on Jeopardy tonight.

 
Comments Off

Posted by on February 15, 2011 in artificial intelligence, DARPA, robots

 

Is the Singularity near? Yes, it is

The very last person I interviewed as a full-time staffer at CBC was a good one. The average person may not have heard of Ray Kurzweil, but in nerd circles, he’s pretty much a guru. Literally. People have referred to him as the “high priest of the Singularity.”

Kurzweil is an American inventor, author and “futurist,” which means he’s famous for making predictions about the future. Not too much unlike the sort of thing a science-fiction author might do.

In our interview, which is up on CBC, we talked about one of his latest projects - Blio, which is an e-reading application that you can download for your computer (and Apple & Android devices soon) that preserves the formatting of the original book. That means all those really nice graphical books, like cookbooks, travel books etc., will look the same on your electronic device as they do on paper.

The meat of our conversation, however, centred on the stuff Kurzweil is more known for - namely, his predictions of the Singularity and a coming future that will almost literally blow our minds.

The Singularity isn’t exactly easy to explain, but it essentially refers to a point in the near future where computer intelligence meets and surpasses the level of human intelligence. The two will merge, Kurzweil predicts, to form a super-intelligence that will make us capable of things we can only dream of right now. This will include making many science-fiction ideas real, like immortality and deep-space travel. A big part of this super-intelligence will come from reverse engineering the human brain, including figuring out how emotions work, which he predicts will happen by 2029. Here’s a video of him explaining it:

Not surprisingly, Kurzweil has his share of critics, who believe he’s smoking the crack. Some brain scientists, especially, say he knows nothing about how the brain works and that we will probably never understand it fully. Predictions about being able to replicate our entire personality into a computer file, which could then live on in a robot or virtual world (hello Battlestar Galactica and Caprica!) are way off base, they say.

The thing I like about how Kurzweil approaches his predictions is that he bases them on something he calls the “law of accelerating returns,” which quantifies the exponential growth of technology over time. I think anyone who covers technology eventually comes to this conclusion on his or her own - I certainly did - that the speed at which new technology becomes available is increasing. This is because if someone over here invents Technology A and someone over there invents Technology B, those are both pretty neat inventions. But when you put them together, you obviously get Technology C, and perhaps D and E and F, and so on.

Technology therefore stacks upon itself, which is why it seems like there are more and more new discoveries and gadgets unveiled every day. It’s not an illusion or an accident - there are more and more every day.

Kurzweil brought up an excellent example in our interview. When the Human Genome Project was started in 1990, people weren’t very optimistic that it would ever get done because so little was known. Lo and behold, the project ultimately took only 10 years to complete, surprising everyone. As Kurzweil explains:

People thought [the Human Genome Project] was crazy in 1990 because only 1/10,000 of the genome had been sequenced by that time. But it kept doubling every year. Half way through the project only one per cent had been collected so the skeptics were going strong, but that was actually right on schedule. One per cent is only seven doublings from 100 per cent.

The other observation I’ve come to is that scientists, while often incredibly intelligent (far more so than me), are generally quite myopic and conservative in their views. They’re afraid of or unwilling to make predictions about where their work can lead, which is pretty much why science-fiction authors exists. Someone’s got to do that job, after all.

It’s also one of the ways in which Kurzweil counters his critics: “A scientist may be sophisticated in his own field but he may not have studied technology progression and he may just apply his linear intuition to his own work.”

Ultimately, those two facts - the exponential growth of technology and the often narrow view of scientists - is why I tend to agree with Kurzweil’s predictions. I recommend reading the interview and if you really want to have your mind blown, check out his latest book The Singularity is Near.

 

A lament for Caprica

Yesterday was an unusual day in that I got to co-host a TV show. I’ve done television before, usually as a guest, but this was the first time I was in a co-hosting role, as it were. The show is Inner Space, on tonight on the Space Channel (11 pm Eastern time, I believe) following Caprica. Check it out, it’s going to be a blast!

The format of the show usually has regular hosts Ajay Fry and Teddy Wilson talking about whatever the preceding program was, followed by a roundup of news from the sci-fi world. Teddy is on an ultra-secret project in an exotic location, so I got the call from producer Mark Askwith to fill in. I’ve known Mark - who I like to think of as the godfather of Canadian science-fiction because of his encyclopedic knowledge and vast contact base - for years, since my days running the short-lived Realms magazine. He helped us out quite a bit in those days, and he’s one excellent dude in general.

We’ve been trying to get something together in relation to Caprica ever since Sex, Bombs and Burgers came out back in March. In the book, I talk about how the themes of Caprica and its precursor, Battlestar Galactica, are actually rooted in reality - particularly military and porn technology - so it seemed like a good fit with the show. I’m pleased we finally managed to make it work.

Alas, I was saddened to hear that the axe has fallen - for now - on Caprica. SyFy, the network that produces the show, announced a few weeks ago that it was not being renewed. Mark and company reminded me yesterday that that’s not the same as being cancelled and there are petitions going around to save Caprica, so ultimately, you never know. Perhaps like some of the characters on the show, it too will be resurrected. (Interestingly, while Space is airing the final three episodes, SyFy hasn’t yet committed to showing them. Translation: there’s a whole lot of Americans Bittorrenting our Canadian-aired episodes.)

If you’re not a fan of either BSG or Caprica, here’s a quick synopsis. In BSG, the human race has created a race of subservient robots called the Cylons. Naturally, the Cylons rebel and nearly exterminate the humans, the remainder of whom flee into space aboard ships. The show was utterly gripping for a number of reasons: it had great drama, suspense and action and it was also a fantastic example of high-concept science-fiction, the kind that looks at modern-day issues through the detachment of a futuristic lens. BSG garnered all sorts of praise - Time magazine even named it the best show on television in 2005 - largely because it asked all the right questions about technology, religion, the nature of humanity, and even terrorism. Yes indeed, it was one of the best shows of all time, if you ask me.

Caprica was/is the prequel series, intended to tell the story of how the Cylons came to be. I fell in love with the pilot episode where Zoe, the main teenage protagonist, explains to her father how she had managed to create a perfect replica of herself in cyberspace using all of the electronic data - email, photos, bank records, shopping receipts, etc. - she had accumulated in her life. I saw this pilot at exactly the right time, just after I’d had some Google executives explain to me how their translation algorithms work, and I saw the correlation that what was going on in Caprica was actually happening in the real world.

Indeed, we are in the early stages of learning how the brain and personalities work and I’m confident that we’ll figure them out, and I’m willing to bet it’ll happen sooner rather than later. Once we get there, we’ll be able to replicate and thereby inject them into the sorts of virtual worlds and robot bodies found on Caprica. And, as the show has been building to, once we can recreate and copy our brains/personalities, immortality will be achieved - we’ll be able to live forever, either in a Matrix-like virtual reality or in a robot body.

I talked about some of this with Magda Apanowicz, the Canadian actress who plays Lacy Rand on the show, when she was in town earlier this year. Here’s the video:

I found Caprica to be at its best when it dealt with these near-future issues. Unfortunately, the show was also tremendously inconsistent, with characters seeming to switch allegiances and/or motivations almost episode-to-episode. Unlike BSG, there also weren’t really any characters who the viewer could like or associate with. There was also almost no humour (unlike Gaius Baltar on BSG, whose self-preservation-at-all-costs and insatiable sex drive were often hilarious, usually tragic), and it was hard to tell in general where the show was going, which is strange for any sort of a prequel because the end destination is ultimately always known.

Nevertheless, I’m hopeful that some sort of miracle happens and that Caprica is saved from death. The show deserves another shot, if only because like BSG, it was asking all the right questions about the trajectory modern-day technology is taking.

In any event, be sure to check out Inner Space. I’m not sure if I’ll be able to get video of it, but if I do I’ll be sure to post it here.

 
1 Comment

Posted by on November 16, 2010 in artificial intelligence, bsg, robots, sex, war

 
 
Follow

Get every new post delivered to your Inbox.

Join 2,897 other followers