Fears of the rise of robots swooping in to steal jobs abound, stoked by the confluence of artificial intelligence, machine learning and other technologies that automate routine tasks traditionally performed by humans. The reality is that job replacement will happen more gradually than the media headlines suggest, and eventually AI will create more jobs than it eliminates, say Gartner researchers.\n"I try to step back and wait for the boomerang because there always will be one," says Gartner analyst Craig Roth, of the hype surrounding job cannibalization via automation. "Once it goes too far to one side, people wake up and realize, 'you know, maybe it's not that bad.'"\nMake no mistake, AI and robotics will have a profound impact on the job market long term. As many as 39 to 73 million jobs will be usurped by automation, according to research McKinsey Global Institute released in November. Yet the consultancy also says that workers will perform new tasks or reskill as tasks become automated, leading to millions of new jobs. Gartner estimates that by 2020 AI will create more jobs than it eliminates \u2014 2.3 million to 1.8 million. By 2025, job creation related to AI will top 2 million net-new jobs, including new positions of highly skilled, management and even the entry-level and low-skilled variety, Gartner says.\n[ Cut through the hype with our practical guide to machine learning in business and find out whether your organization is truly ready for taking on artificial intelligence projects. | Get an inside look at 3 machine learning success stories. | Get the latest insights with our CIO Daily newsletter. ]\nSuch statistics contradict the panic-inducing headlines that millions of jobs will evaporate. One reason for this is that people tend to conflate job automation with automation of certain tasks, which is a much more accurate snapshot of the current corporate climate, Roth says. In other words, robots and other machines aren't replacing humans outright, but sharing the workload.\nRobots and humans making pizza together\nTake a pizza business that employs robots to work alongside humans. Maybe the robots perform roughly 10 of 20 tasks it takes to build, bake and deliver a pizza, says Roth. In fact, such a business already exists in Zume Pizza, a Mountain View, Calif.-based startup that uses several robots to help humans assemble and bake pizzas.\u00a0\nHumans prepare and divide the pizza dough, put it into a dough press and then onto a moving assembly line. When orders come in, one robot applies sauce and another spreads the sauce before handing it off to a human to smooth out imperfections, says Zume CTO Josh Goldberg. The human worker then adds cheese and toppings, which requires fine-motor skills robots are, for now, largely incapable of. "Humans do it today because it's a pretty highly skilled thing to do," Goldberg says.\nAnother robot takes the finished pie and places it into a par-bake oven, which heats the pizza at 800 degrees to kill the active yeast in the dough before transferring it to trucks, Goldberg says. Beyond saving humans from burns, Zume's value proposition is fresh, farm-to-oven ingredients and speed. Zume is using trucks equipped with "bake-on-the-way" technology that begins cooking pies shortly after they are ordered; the average delivery time is around 22 minutes.\u00a0\nZume uses optical targeting and computer vision technology to inform the automation for the base robots, which are made by ABB, and has developed custom tooling to enable the robot's manual operations. Roth says Zume offers a glimpse of the human-machine labor division typical today.\nWhat the future of AI means for work\nYou can think of AI, robotics and even robotic process automation (RPA) as endowing humans with certain "superpowers," enabling them to get more tasks accomplished in less time, Roth says. Gartner calls this "AI augmentation," and posits that one in five workers engaged in mostly non-routine tasks will rely on AI to do a job by 2022.\nThe growing application of AI technology to automate tasks also means IT leaders must take into consideration what jobs will be lost, what jobs will be created, and how staff will collaborate and get work done. CIOs may have to reskill employees and reposition them for other services that provide greater value for customers, Roth says. Ultimately, CIOs must prepare their organizations for an automation boom that will trigger significant, though hard to quantify, changes.\nTo that end, Roth advises:\n\nAvoid hype. Don't get distracted by shiny new AI technologies or dread-inspiring headlines that say robots are taking over the world.\nWatch what you say. When talking about the automation capabilities of a certain technology, don't hammer home the fact that it could lead to headcount reduction. Maintaining a low-key, sober approach is crucial in conversations with employees. You don\u2019t need to incite a panic.\nKeep societal impacts out of the equation. Consider how AI and other automation capabilities can fortify the business and drive competitive advantage, but don\u2019t worry about the societal impact these technologies portend. Figuring out the right technology decisions for the business is hard enough without falling into a moral quandary. That's for academics to worry about, Roth says.\nApproach AI lightly. Given their disruptive potential, don\u2019t rush into AI and automation.\nAvoid the \u201cnuclear AI\u201d problem. Another reason to proceed with caution is the potential for something to go awry, leading to a public relations disaster that sets the AI and automation markets back years. For example, Microsoft didn\u2019t do the AI software sector any favors when it\u2019s chatbot Tay unleashed racist, sexist and other incendiary comments on social media. One possible scenario: What if a self-driving delivery car operated by Zume Pizza gets into a serious car accident? \u201cThere's a risk that some employers make stupid decisions about AI that could cause a backlash,\u201d Roth says. \u201cThere are so many opportunities for that to happen.\u201d\n\nBottom line: \u201cIf [CIOs] can sprinkle [AI and automation] in lightly and in a humanistic way over time, that will do a lot more for society than trying to solve the challenge at the societal level,\u201d Roth says. \u201cAnd if a CEO or CIO could convince their peers to follow suit, that will have the most impact.\u201d\nRelated artificial intelligence articles:\n\n\n A practical guide to machine learning in business \n Machine learning success stories: An inside look \n 9 IT projects primed for machine learning \n 10 signs you\u2019re ready for AI -- but might not succeed \n 10 strategic tips for getting started with machine learning \n Which deep learning network is best for you? \n How to build a highly effective AI team \n Why you should invest in AI talent now \n Why AI careers can start with a degree in linguistics \n The year of Alexa and the coming decade of A.I.