Artificial intelligence research is being carried out on multiple fronts, by neuroscientists and computer engineers studying everything from stem cells to quantum physics.
Over the next 20 years, we'll likely see technological breakthroughs in hardware, software and wetware that will change everything in an instant, sending research on a totally new trajectory.
But experts contend we can still make some educated guesses about how AI will drive social and economic changes by the year 2035.
"Just as the Industrial Revolution changed every aspect of society, the A.I. Revolution will have the same impact," says Christopher Martinez, professor of electrical and computer engineering at University of New Haven. "The Industrial Revolution changed manufacturing and the A.I. Revolution will change the intellectual landscape."
Martinez cites the case of IBM's famous A.I. project Watson, which has evolved from a TV game show contestant to a sophisticated cognitive computing system now involved in complex cancer research. "Instead of playing a game of Jeopardy, computers similar to Watson will change health care, marketing, education, service industries -- almost every occupation will be using A.I.," Martinez says.
Tom Davenport, professor of information technology at Babson College and a research fellow at the MIT Center for Digital Business, recently co-authored a report for Harvard Business Review concerning the future of A.I. on jobs and economies. Davenport's take: A.I. will have a radical impact in the workplace by the year 2035 -- but we humans are notoriously good at adapting to new developments.
"I think there is cause for concern," Davenport says. "Increasing numbers of knowledge workers will have at least some substantial component of the their jobs taken over. Any particular narrow task, in the future, we can probably do it faster and better with machines. So these narrow tasks will be taken away, one by one, but there will still be things that humans do better."
In the article, Davenport and co-author Julia Kirby outline several different ways in which the threat of automation can be turned into an opportunity for augmentation -- A.I. and people working in collaboration.
"There will be jobs for big picture thinking, integrative thinking, tasks with emotional components or creative aspects," he says.
"When I look into the past and think about the future, there were periods where great numbers of workers were displaced. But the people who actually bothered to learn how the machines worked -- to find a role configuring them, fixing them, monitoring them -- they did quite well. I think the same thing is going to happen with our smart machines."
Of course, assuming we do have true A.I. by 2035 -- actual sentient, self-aware machines -- there are broader and weirder societal implications as well. Will artificial intelligences have citizenship? What will be their rights under the law?
Would they be allowed, as corporations currently are, to spend money in political campaigns? Will they vote, and maybe more importantly, how will they vote?
Jenny L. Davis teaches sociology at James Madison University and specializes in the intersections of technology and social theory. Davis says she believes we will have sentient machines by 2035, and that things are going to get complicated.
"Technology always traces back to its human creators and users, and therefore reflects the power hierarchies within which they were built," says Davis, who co-edits the wide-ranging Cyborgology blog.
"The concern is that artificial intelligence will reflect the logic of already advantaged groups -- white, upper-middle class, highly educated, from developed countries -- while excluding those who are already disadvantaged. As technologies develop under the auspices of objectivity, it's easy to lose sight of how those technologies are deeply value-laden."
Davis says she fully expects to be living and working alongside some incarnation of true artificial intelligence by the year 2035.
"What is less clear is if we will -- or even can -- have machines that are emotionally aware. Humans can program emotional responses, but will machines truly feel? If they can, are they still machines, or something else entirely, something more human?"