The conversation about artificial intelligence is downright mythological.

In mythology and fairy tales, we encounter all kinds of strange mixed-up beasts: griffins that are part lion and part eagle; Pegasus, the horse with wings; and centaurs, noble creatures with the torso of a man and the body of a horse. These figures combined the strengths of their constituents and consequently became more than the sum of their parts. Fearsome and sometimes misunderstood, these chimera were a way of making sense of a changing and often troubling world.

But what does this have to do with artificial intelligence? AI is now telling the modern version of these mythological stories. Like centaurs, AI’s place in our current narratives is moving from that of a misunderstood but powerful beast to guiding us through a changing world and making us better humans along the way.

In AI there’s something called “the centaur effect,” a vision of seamless integration named for the half-man, half-horse creature. The centaur effect represents a more exciting and nuanced view of AI, in which a human-machine partnership will improve existing jobs and create new ones that we can’t yet imagine. The centaur effect imagines AI-powered machines taking over much of the rote and repetitive work of current jobs — often with a voice-controlled assistantallowing humans to focus on higher-level tasks that require social skills, empathy, and problem-solving. Nicola Morini, Global Managing Director for AI at Accenture, sees how AI will change customer service by empowering its workers.

“In the B2C world, you don’t want to wait 15 minutes on hold to pay a bill,” he said. With an AI system handling the basic tasks, humans are free to handle more complicated interactions. “Humans are happy because they’re doing higher-end jobs, and for [the customer], the interaction happens much faster.”


A Smarter Beast


AI serves us better because it understands our needs better. Five years ago, basic tasks were so labor-intensive that there wasn’t a practical application for AI, said Richard Socher, Chief Scientist at Salesforce. Now, a customer can send an email or chat to a business and AI can immediately classify it, based on its content, as intended for billing, support, or sales. “As we make more progress in AI, we’ll see that shift from research to production,” he said.

Morini added that the business process itself will change. In the past, people would write a program for a computer to solve a particular problem. With AI, he said, people will design the general constraints of the problem, “then we’ll let the machine decide what is the best path to take to optimize that particular process.” We’ll set the path that the machine will use to take us places we’ve never been before.


Where Centaurs May Roam


By helping us work faster and more efficiently, AI makes us more productive. Socher noted how AI will preserve jobs by enhancing them. “If you make every salesperson 30% more efficient, the company won’t fire 30% of the people,” he said. AI improves the employee’s ability to create profit, thereby ensuring the value of both human and machine.

How AI will change the workforce is the “central question,” said Terah Lyons, former policy advisor to the White House Office of Science and Technology Policy under President Obama. It’s not as simple as looking at pay scale, she said. Truck drivers and housekeepers make about the same money, but while truck driving can be automated, housekeeping requires skills that machines can’t replicate. “I don’t think we have to worry any time soon that these positions will be automated out of the marketplace,” she said. “So you can’t segment by pay scale.”

While working with the White House’s team, they came to some conclusions. Overall, she said, “we did think that AI would probably increase wealth. But we should continue to invest in AI to make sure the benefits are shared.”


Let the Machine Handle the Boring Stuff


Centaurs will be created anywhere AI can augment human workflow. Machines can’t code better than humans, but they can do the testing so the code is analyzed faster. Machines also can’t translate better than humans, but humans can run copy through a translation app and then work off that rough version. AI is elevating the nature of work by enabling humans to focus on the most-human parts of work.

Handling the administrative, the rote, and the boring, AI will take over a significant portion of essential baseline tasks — scheduling service calls, dispatching agents, answering those frequently asked questions that eat up so much employee time. Companies can then retrain service agents as salespeople: They can use those calls that come to them to cross-sell. By blending humans and AI in the workplace, entirely new jobs evolve.

Accenture conducted a survey of 1,000 large companies regarding how they use or plan to use AI and found three emergent categories of new jobs: trainers, explainers, and sustainers. Accenture defined them: “Trainers teach AI systems how to perform, process data, and behave. Explainers articulate to business executives the inner workings of AI systems, improving transparency. Sustainers ensure that AI is fair, safe, and responsible.”


Other New Creatures


In these new jobs, employees will act as intermediaries between AI systems and humans. Explainers will teach executives how AI works so that they can make informed decisions about the use of the technology and sustainers will perform a necessary ethical role in making sure AI doesn’t cross any significant lines in, for example, gathering data on customers. Trainers may have the most interesting jobs in teaching machines how to navigate the tricky terrain of such things as language: detecting sarcasm, for example, in a chat message. Other trainers will be responsible for imparting an understanding of empathy. Humans will always be needed when dealing with customers who are also humans.

From the customer experience to the structure of business processes to the workforce itself, our relationship with AI is unpredictable — but that’s what makes it exciting, as Socher said. When ARPANET, the primitive internet, was switched on in 1969, “nobody would’ve guessed that there’d [someday] be a social media marketing manager job,” he said. “It’s obvious to me that human ingenuity and creativity will prevail, and people will find interesting new things to do.”

Nicola Morini-Bianzino, Terah Lyons and Richard Socher were speaking in a panel titled "AI Demystified: From the Science Lab to the Workplace" that took place at TrailheaDX in June 2017. To see a video of the entire panel, visit Salesforce Live.