Over 150 years ago, British author Samuel Butler predicted the rise of artificial intelligence, calling for a “war to the death” against machines – and arguing that that “the time will come when the machines will hold the real supremacy over the world and its inhabitants.”
Today, the inevitable conflict between man and AI-powered machines permeates our national discourse, as the threat of technological unemployment looms large. Elon Musk recently told a gathering of governors that AI is “the greatest risk we face as a civilization.”
Photo by Chris McGrath/Getty Images
But while the national narrative tends to reflect Butler’s dystopian fears for the 3.5 million truck drivers rendered obsolete by autonomous vehicles, or Mark Zuckerburg’s often mischaracterized vision of an educational future devoid of teachers – early adoption of AI suggests a far more collaborative reality.
Because in practice, AI often shines a light on areas where replacing humans with robots leads to suboptimal results – but cooperation between humans and machines, create outcomes that are better than either might achieve independently. Google Translate results may be technically accurate, for example, but fail to transpose idioms or slang that human translators can interpret. Interactive voice response (IVR) systems, unable to deal with the breadth and complexity of customers’ needs, trap frustrated consumers within the endless computerized loops until a human, armed with information gathered by computers, can direct them toward a solution. And in the education context, outcomes for AI-driven courses have failed to produce results, at scale – without the thoughtful support and encouragement of real-world teachers.
Of course, the concept of human-machine symbiosis isn’t entirely new. Average chess players paired with laptops famously took down chess masters and supercomputers in one 2005 match. PayPal’s founders paired human analysts with sophisticated algorithms to tackle complex fraud challenges – giving birth to the technology that undergirds mercurial tech titan, Palantir.
At American colleges and universities, human-computer symbiosis is allowing faculty like University of Michigan Professor Perry Samson to leverage students’ mobile phones to collect real-time information about student behaviors, and modify instructional strategies to improve learner outcomes. Technology helps teachers understand how students answer problem questions, whether they take notes, get confused, tag content in books, or review lecture material after class. Rather than replace them, technology is making human teachers more effective, at a time when even the most advanced AI isn’t able to data mine its way to non-obvious hypotheses that improve student learning. Educators, empowered by data, are drawing on human intuition and creativity to identify correlations between student behavior and outcomes. Technology isn’t just making teachers more effective, it’s enabling a new era of pedagogical innovation as a growing number of educators experiment with flipped, blended, and adaptive courses.
This approach is not unique to education. In customer service, the failure of IVR to deliver a positive customer experience has led to a blended approach that leverages AI to match customers with the customer service reps most likely to address their content needs – or even personality. Rather than replace call center operators, companies like T-Mobile are using AI platforms (like little-known “unicorn” Afiniti) to drive improved results of existing employees – by simply using AI to learn from historic interactions to better pair customers with their call center representatives. Same team, same customers, but upwards of 5% improvement in sales outcomes and retention with significant impact on the bottom line.
The applications of blended AI are broad. Stanford researchers have used advances in machine learning to develop a human-machine hybrid for translation, allowing bilingual human translators to move faster than they could if translating everything manually, while also improving the accuracy of machine translation. The blended approach allows for quicker translation on the basic language and uses humans to finalize the more subtle portions of the translation for context and culture.
A final dimension that most futurists are missing in discussing AI is that it actually creates whole new categories of jobs in training AI technology, explaining the contextual situations that AI machines don’t handle well (sarcasm) and measuring efficacy. Just as computers never really eliminated paper, we realize that AI – like the internet and renewable energy – will create whole new career paths.
In each case, AI is transforming the way humans interact with each other, in ways that make those interactions more efficient, effective, and even more meaningful. Could it be that AI might actually facilitate more authentic connections between individuals? The robots may still be coming for our jobs. But rather than competing with people, AI may turn us into the real supercomputers – and, in an ironic twist, make human interaction even more human