Connecting artificial intelligence with the internet of things

It’s no secret tech luminaries like Elon Musk and Bill Gates worry about humanity flirting with disaster though a digital version of the Icarus myth – in our case, the power of artificial intelligence being the sun that eventually burns our wings.

Even so, not every futurist or technology pioneer is quite so alarmed that our experiments with AI might confine us all to dystopian doom.

John Underkoffler – the chief executive of Oblong Industries, better known as the guy who created the futuristic gesture-based interface in the movie Minority Report – is one such thinker. He founded Oblong Industries in 2006 as a first step to bringing ideas like the Minority Report interface into the real world, which means he knows as well as anyone where the bounds of sci-fi end and a high-tech new reality begins.

Tom Cruise as detective John Anderton in Minority Report. Facebook Twitter Pinterest
Precrime fighter … Tom Cruise as detective John Anderton in Minority Report. Photograph: AP
His company today sells what it describes as commercial versions of the Minority Report computers.

Relying on them in the conference room, using a connected “wand” to manipulate on-screen data, can make the user look almost like they’re conducting an unseen symphony. That’s one manifestation of how technology giant Cisco has projected that the so-called internet of things (IoT) – the networked connection of people, process, data and things – will see connections surge up to as many as 50bn by the end of the decade.

Reasons for optimism

Those connections represent a whole new universe of devices coming online, sending out data, pinging other devices and servers – and, depending on who you ask, eroding more of our privacy in the process, or worse. Yet, as Underkoffler sees it, there’s no need to fear the machine-filled future that awaits us, at least not yet.

Instead, he thinks there are some basic questions that need answering first, before it makes sense to even begin to think about putting limits around potential AI capabilities.

“The optimist in me says, okay, millions of new objects all connected to the internet – wow, to make sense of that is going to require an incredible new interface,” he says. “I’d love to start thinking about that. How do we talk to all these objects in a coherent way? That’s a really great design problem.”

He and some of his colleagues from around the globe believe an over-abundance of concern at this point about a super-intelligent AI running amok, co-opting the IoT and turning our gadgets against us means we’re worried about the wrong things – and not asking the right questions.

Professor Sanjay Sarma – the director of digital learning at MIT and a pioneer who helped develop the technical concepts and standards behind radio frequency identification technology (RFID) – said he believes that, on balance, AI will deliver value as we gradually connect our thermostats, refrigerators and the like to the web, and to each other.

He also points to a potential bogeyman on the horizon, one that he thinks trumps potential AI-related mischief. It’s his fear that the introduction of an entirely new device class, one for which a prevailing digital architecture does not yet exist, introduces all kinds of potential new security vulnerabilities that flourish in the ecosystem’s gaps.

“All technologies converge,” Sarma said. “It’s inevitable. The benefits, in my view, are potentially incredible regarding the IoT. The BP disaster – imagine if AI were watching over those systems and could have detected the disaster earlier.

Potential pitfalls

Sarma continues: “Do I have worries? I do. I’m more worried about artificial stupidity. I’m less worried about systems so intelligent they out-do human beings. I worry we will build artificial intelligence systems that are too smart by half, where they do something really dumb – for example, a cascading series of events that results in a power shutdown. Or a poorly designed system that gets hacked and causes havoc. I’m more concerned we will create flawed systems which compromise our privacy and our security, even unintentionally.”

Indeed, Sarma thinks it’s only a matter of time before some poorly designed IoT system is hacked by a major player, creating havoc on the scale of something like a large area power shutdown.

One reason? The existence of so many “walled gardens” in the sector.

“I have several Nest thermostats in my house,” Sarma said. “They do one thing and do them well. But what if I want to buy a Nest system and an internet-enabled home lock, and I want to bridge the two? Where, say, I unlock the door and Nest increases the temperature. And I want to connect them to my Tesla. Well, because they’re walled gardens, what’ll happen is someone somewhere is going to do something stupid – some vendor will bridge it together – and doing that will open up security and privacy holes.

“There’s a huge opportunity here that’s inevitable, and we’re all barreling towards it. But all great systems are based on clear, simple to understand architecture. The world wide web is based on hyperlinks. We don’t have a clear architectural understanding of what the IoT is. Everything is walled gardens with bandaids on top. We need to have one single garden of eden we all play in.”

Meanwhile, count the technologist who envisioned Tom Cruise’s character in Minority Report swiping through data among those who think our present fears about AI-IoT are overblown.

“I’m perplexed by that one,” Underkoffler says about warnings like those from Musk, who earlier this year donated $10m to the Future of Life Institute to finance research into keeping AI under control. “They should honestly know better. They’re either badly informed or irresponsible to be fear-mongering in that way, because the truth is we don’t have AI.

“We should distinguish between AI and machine learning. There’s a lot of debate about the difference, but I think we can distill it down to consciousness. The decision to manipulate the surrounding world–- machine learning systems don’t have the slightest bit of that. The systems we build today are not built with a mechanism of modifying themselves. The mail sorting machine is never going to decide to turn into a genocide machine. The span of functionality never changes.”

Shades of grey

Ian Pearson, a UK-based future technology consultant with Futurizon, feels more or less the same. He also sees the IoT as representing significant good as well as something for which customer trust could easily be broken.

“I think if you could guarantee our privacy, most of us would love to have an environment like a smart home that adapts to what you want – lights that dim according to your emotional state, things like that could be a very nice environment,” he says. “Right now, a lot of what we have by way of AI is hopelessly rubbish. If it’s nice to us, bring it on. If we can get the benefits of AI and connectedness and it’s absolutely benign, I think it will be fantastic.”

Underkoffler, whose career has included consulting on sci-fi Hollywood productions like Iron Man and Hulk in addition to Minority Report, also concedes there could be a dark side to all this. But he and futurists like him think the good and bad of bringing true AI to the IoT sphere won’t matter until other things happen first, like making sense of the jumble of user interfaces and the patchwork of code and security frameworks currently represented by all our connected devices.

“My question is how you construct the user interface – what is it for? What does it do?” Underkoffler said. “There’s a question around what would be a central, universal UI, like an atom we could attach to every object to indicate, let’s say, if it’s taking information it’s collecting back to some location. Some equivalent to a dial tone – the basic bit of information that tells you something is working in a particular way.

“After that, it seems like we’d need at least a minimal Bill of Rights, as it were, that we could then attach to the IoT.”

“I really do think that people owe themselves a dialogue about technology that steps aside for one moment from the kind of constant churn of consumerism – away from whether you have the latest iPhone and what about the Apple Watch and so forth – and takes a look at the whole picture,” Underkoffler says. “Where’s the value? And who gets to be in control of that? Who gets to design and who gets to build it?

“Of course, the moment you start asking those questions, the answer is evident – you do. if you can ask a question, you can answer it as well. And that’s a healthier world to live in. Where people have agency and can take it on themselves to modify the world around them.”

Credit: illustration by Ian Pearson Photograph: Xinhua/Alamy

Full article available at The Guardian