A quick recap of the best moments of the Roundtable. (more)
Kevin Kelly outlines the big issues in play in this Roundtable. (more)
A short teaser video of the roundtable – perfect for sharing. (more)
For those who think that Artificial Intelligence is the stuff of science fiction in some distant future, think again.
This fascinating roundtable conversation among a top group of folks involved in creating or tracking AI clearly showed that this fundamental enabling technology is already here, more is coming fast, and the ramifications will be large and transformative.
The conversation built through successive looks at the short-term, medium-term and long-term picture for AI and each stage became more exciting and controversial. Exciting, in that our children and grandchildren may well live in a “fairy land” where almost all the previously inanimate objects around us will talk to us and have some form of intelligence. But controversial, in that the transition to that world will displace many current jobs requiring knowledge and human intelligence, and will stretch the bounds of our current laws, institutions, and even ethics.
The best thing you can do is watch the Roundtable Recap video that pulls together in about 10 minutes all the best insights from the fully 90 minutes with a thread of my analysis. In general it follows the sections below.
Short-term: The Future is Here
We came to the roundtable with a focal question to look at what will be the early path of commercialization for Artificial Intelligence. Very quickly the participants started pointing out that we already have early forms of AI in the commercial space like with NetFlix, which learns the kinds of movies that you like.
But then those at the roundtable began talking about actual products that are not as well known that can be described as using AI. Like high frequency trading. Or George Dyson talked about military drones. Or Nova Spivack talked about some of the current powerful computers that are sifting through big data and starting to really learn, which is the key attribute of intelligence.
Krisztina Holly talked about a product she knows of that is still in the lab that helps elderly people in physical workouts and learns their preferences and how best to deal with them.
Medium-term: AI as a cloud service
If the key is to learn, then AI would operate best connected to large and ever-expanding databases. In other words, it would probably inhabit the cloud.
Kevin Kelly made the case that AI might become a cloud-based service that you would rent or lease when you needed it. If that were the case, then there would be advantages to being an early mover and grow through the network effect – making it hard for new companies to catch up.
Doug Lenat took the next step and said that eventually AI would be like electricity and be found connected to everything. This is the long-held goal of ubiquitous computing where even the doors can talk to you and know what doors need to know.
Long-term: weak telepathy and super smart machines
It kept getting stranger the longer time horizon we talked about. Doug talked further about how machines will learn so much about us that they will be able to detect the slightest suggestion of our wishes and give us what we wanted – almost as if the machine was telepathic. But it will be weak telepathy and it will only be reading super-subtle signals.
And, of course, as in any conversation about AI, there was the specter of what happens when the machines learn so much so fast that they get smarter than humans.
A good chunk of the discussion did talk about how we might deal with the ramifications. Kevin talked about researchers steering clear of full artificial intelligence and focusing more on artificial smartness. Keep machines focused on doing discreet tasks extremely well and do not strive to make them conscious. He called it domesticated AI, or even castrated AI.
And there was a lot of discussion on developing an ethics sooner rather than later in this AI space. In fact, in the ethics discussion, some thought the machines, always ones to follow rules, might be more ethical than humans in the end.
It was a strange conversation, but an important one. Watch it for yourself. Full Post →
After decades of broken promises in the field of Artificial Intelligence, we’re seen some big advances in recent few years. A new AI era is about to arrive. Siri and Google Voice are early indicators of a broad sweep of things that could be done by nascent, and ever-improving, machine intelligence.
AI is an enabling technology that can be expected to have big implications in many fields, from robotics to automation to manufacturing. But AI will go well beyond fields of manual labor and enter the realm of knowledge work so the repercussions on the entire economy and national employment situation will almost certainly be profound.
Talk about reinventions. The arrival of AI could reinvent our role in the world, what it is to be human, to be employed. As we shift many kinds of work to machines, it might ultimately require our society to rethink full employment, the hours in a work week, the social compact among citizens.
Few people are better suited to think through AI’s possible trajectories and implications than Kevin Kelly, the founding executive editor of Wired magazine and the author of many important books on new technologies, including What Technology Wants, and the one he’s currently writing called “Technium Rising.”
Kevin will anchor this Reinventors Roundtable, the first of this season’s Reinvent America season, and he wants to focus on this key question: What is the early path of commercialization for Artificial Intelligence? In other words, what will be AI’s first killer apps?
Kevin thinks it’s important to start there because he wants to see what happens when AI gets out of the labs and regular people start to engage and inevitably evolve it. The technology will only really matter once there are viable businesses to sustain its development and expansion. So where should we be looking?
Kevin has an hypothesis that the arrival of AI will be in the form of thousands of different discrete and focused applications of what he calls Artificial Smartness. Machines will get really, really smart at doing specific things better than humans or that humans simply could not do. However, these machines will not need to be conscious, in the way human intelligence works. In fact, we might actively avoid going down that path of development.
Reinventors will be bring together an eclectic group of experts, entrepreneurs and other unusual thinkers to work on this important and complex problem. We also expect to think about what could, or possibly should, accelerate or slow down these developments too.
Kevin is generally positive about the possibilities of what is to come with AI, but at the very least he thinks we would be well served to prepare for its arrival. How should humans intelligently embrace artificial intelligence? Let’s get a handle on this powerful enabling technology right now. Full Post →