Watson is evolving – or better said, his children and offspring are thriving. Watson himself will be there, somewhere in the background, a giant ultra-fast database and immensely clever algorithm that can be consulted as desired. From AI to AI, so to speak. Somehow this keeps reminding me of that magnificent Azimov story, “the final question” with the omnipresent, unbelievably powerful “AC” (ultimately a galaxy-wide network of colossal AI computers) taking care of stuff behind the scenes. Are we going that way? Well, yes, and no.
Daddy, you have a flame-out in your port engine
Many, many moons ago – more moons than I care to remember – in a galaxy far, far away I was a programmer. Not a remarkable one, nor famous, nor rich, but I enjoyed it – still do, on occasion – and I did, to my (admittedly, odd) taste, some interesting stuff, especially a decade later with the then-emerging 4GL database-enabled languages, compilers and tools. Combine that with beginnings of – by then – reasonably powerful and useful AI technology and you have a potentially interesting field of endeavor. It’s a long and colorful story.
Around that time – let’s say – “some militaries in some countries capable enough” were also looking into AI systems for military applications, such as pilot-aircraft avionics control systems interfacing, where the human-machine control interface was such that it allowed the pilot to concentrate to the important stuff – such as where to point the airplane – whilst the AI took care of many other equally important “background” tasks. Or, co-pilot tasks, if you will. There were already some rudimentary speech interfaces. One of the breakthroughs was the fact that some cool stuff in the form of AI-human integration was developed. Primitive by today’s standards, no doubt, but interesting. For instance, the AI would talk to the pilot in the voice of his daughter or wife, and inform him of corrective or preventive actions taken (“I just deployed more flares”) or dangers, or worse. This turned out to be very effective. Instead of a cacophony of shrill alarm signals and flashing indicators, potentially completely overwhelming the pilot or slowing him down trying to digest it all, there would be a voice: “Daddy, you have a flame-out in your port engine” and this would immediately get the right amount of attention and action.
I have no idea what the current state of affairs is at this point in time, as I am no longer involved in that kind of work. But I can imagine that it is already lightyears ahead of those fascinating pioneering days. Judging by what is available these days in terms of event management / incident management tools, and event / incident correlation engines for service management software, it’s likely quite advanced. Interesting also is what’s going on with Watson and his – not yet so famous – offspring. Yes, Watson has been naughty. There’s many mini-Watsons now. Miniaturized, and therefore not as powerful, but quite smart – and also interfacing with Watson himself or some of his other close and not-so-close family members.
The ultimate intermediary
Just as it easy to overwhelm the pilot of a complex combat aircraft with event data, alerts, and signals, it is possible to completely overwhelm the customer using his smartphone with a myriad of passwords, logins, and similar actions required to get something done. Integration is needed – a intermediary. Therefore, the goal is not just to build great artificial intelligence. Some companies see in this effort the opportunity to become the ultimate intermediary between businesses and their customers. Search engines were among the first of these “platforms,” enabling Google to generate a fortune from organizing the vast array of Web pages and databases for ordinary users. Then, with the rise of smartphones came apps that pulled consumers out of desktop search into the mobile world. Apple and Google raced to become the gatekeepers of these smartphone programs by building app stores that take a cut of the profits.
Waning enthusiasm for Apps
But despite apps growing into a remarkable $50 billion (!) business, consumer enthusiasm for most new apps is waning, according to some analytics companies. Too many apps, none of which speak to each other, living inside their own limited world, trapped in between the walls of the gardens of apps stores owned by Apple and Google. Or, at least, that is how many see the current multiverse of Apps. The result is an ineffective array of non-coherent, non-consistent tools – too much data used, too much bandwidth, too many passwords to remember, too many useless notifications.
Mobile users now spend 80 percent of their time in just five apps, according to 2015 data from Forrester. It’s just too inconvenient, too complex, and too time-consuming for consumers to hop in and out of so many apps. So, naturally, consumers are determining what apps are the most productive and effective to them, and barely use the rest. If you’re a programmer – and an app builder – there’s now a much bigger bar to get over if you’re going to build a new app, just to get users to take note and use your app.
In the beginning
In the beginning, there was Siri – the Apple assistant app. Siri isn’t a real AI, but capable. In fact, the original Siri – launched by now quite some years ago – was in many ways more capable than the current Siri. The original Siri wasn’t supposed to be a clever AI chatbot. The goal was to reinvent mobile commerce itself. When it initially launched as an independent app in 2010, Siri could buy tickets, reserve tables and summon a taxi — all the while bypassing search pages and without a user having to open or download another app. She was able to siphon data from 42 Web services, including Yelp, StubHub, OpenTable and Google Maps.
But nearly all of the partnerships were dissolved once Apple took over for reasons only clear to Apple. The original developers of Siri left Apple later that year.
The stealthy, by now four-year-old Viv is among the furthest along in an endeavor that many in Silicon Valley believe heralds that next big shift in computing — and digital commerce itself. Over the next five years, some believe that transition will turn smartphones — and perhaps smart homes and cars and other devices — into virtual assistants with supercharged conversational capabilities.
Powered by artificial intelligence and unprecedented volumes of data, they could become the portal through which billions of people connect to every service and business on the Internet. It’s a world in which you can order a taxi, make a restaurant reservation and buy movie tickets in one long unbroken conversation — no more typing, searching or even clicking. The data from these services enables the Viv brain to seem “intelligent.”
But Viv is by no means alone in this effort. The quest to define the next generation of artificial-intelligence technology has sparked an arms race among the five major tech giants: Apple, Google, Microsoft, Facebook and Amazon.com have all announced major investments in virtual-assistant software over the past year.
Virtual assistants offer an alternative to the myriad of apps. But the difficulty, stemming back to the early artificial-intelligence efforts in the 1960s, has always been understanding the nuances of how humans talk. Most virtual assistants today can understand a set of human questions. But those queries have to be stated in a precise way, and they trigger largely scripted responses. What distinguishes Viv is that it aims to mimic the spontaneity and knowledge base of a human assistant.
By working with data from movie-ticket vendors, it can understand the multitude of ways people can ask it to buy movie tickets. It can look up showtimes and, on its own, suggest entertainment alternatives from other vendors if the desired showing is sold out. And it can compare prices and then buy the tickets, along with making a restaurant reservation beforehand. If the user changes her mind, the assistant can take care of the cancellations and let her know it’s done.
We’re going to see a lot of interesting developments in the coming years, that’s for sure. In fact the developments in AI are going so fast that many – even experts in the field – have trouble keeping up. The exact effects and workings of AI technologies are becoming more challenging to perceive and comprehend for humans. Effectively, as the effectiveness and impact of AI technologies increases, the more difficult to understand their working and impact. It is easy to see and predict that in the near future artificial intelligence may become more intangible, indistinguishable and, ultimately, quite incomprehensible for humans. This means that AI technologies will soon go beyond Clarke’s third law, stating that “any sufficiently advanced technology is indistinguishable from magic.”
As various incarnations of Watson already have demonstrated – and quite impressively – todays various machine learning systems can already provide unexpected insights in all kinds of fields, from personalization technologies, biology and particle physics, from cooking recipes and outlandish game moves to crime prevention. And that’s only for starters.
AI technologies could evolve into a back-office-like platform, an largely invisible infrastructure similar to the Internet, that would allow people themselves to decide the way they utilize AI or contribute to its design and development. A potential, easy interface to that grid would be systems like Viv. Such an AI grid, accessible through smart interfaces powering various experiences and applications in different environments and industries, being open for tinkerers and specialists alike, would significantly change the way we could understand AI or interact with intelligent systems in general. Human and machine intelligence would be intertwined in unseen ways.
In many ways, this again reminds me of the AC from Azimov.
The border between physical and digital realities is beginning to dissolve. If and when the relationship between humans and intelligent systems gets more seamlessly intertwined, the border between human intelligence and artificial intelligence might begin to dissolve, too.
Come to think of it – as it ultimately did in Azimov’s story.