AI and machine learning becoming popular, is often seen as threat of job replacement rather than as opportunities for job design.
According to MIT professor Erik Brynjolfsson and his colleagues, this debate needs to take a different tone. We are moving towards designing human machine collaboration for better future work.
There is nothing to worry about job losses, executives should be helping to reduce jobs in which AI and machine learning take over boring task, while humans can spend more time with higher level tasks.
Barbara Grosz says “AI systems will need to be smart and to be good teammates”.
New research finds that automation is not all about replacing the entire occupation but the some specific tasks within the jobs. Some jobs will be more heavily impacted than the others.
“Our findings suggest that a shift is needed in the debate about the effects of AI: away from the common focus on full automation of entire jobs and pervasive occupational replacement toward the redesign of jobs and reengineering of business practices,” the researchers write in an article published in May in the American Economic Association Papers and Proceedings.
“Despite what Hollywood is saying, we’re very far from artificial general intelligence. That’s AI that can just do everything a human can,” Brynjolfsson said. “We don’t have anything close to that. We won’t for decades, unless there’s some amazing breakthrough.”
What we do have are powerful narrow AI systems which are capable of solving certain, specific problems at human or super-human levels of accuracy, typically using deep neural networks. Those technologies are adept at tasks involving predictive analytics, speech and image recognition, and natural language processing, among others.
“But that’s not everything—it’s some things,” he said. “That raises the obvious question: what are the tasks that this amazing AI can do well, and which are the tasks they can’t do?”
To answer those questions, the researchers developed a 23-question rubric to determine whether a task is suitable for machine learning. How high or low a task’s score is on the rubric indicates how susceptible it may be to automation and machine learning, Brynjolfsson said. He and Tom Mitchell published the original rubric in the journal Science in December, 2017.
“Automation technologies have historically been the key driver of increased industrial productivity. They have also disrupted employment and the wage structure systematically,” the researchers write. “However, our analysis suggests that machine learning will affect very different parts of the workforce than earlier waves of automation … Machine learning technology can transform many jobs in the economy, but full automation will be less significant than the reengineering of processes and the reorganization of tasks.”
Radiologists, for instance, have 26 distinct tasks associated with their job, Brynjolfsson said. Reading medical images is a task well-suited for machine learning, with computers starting to become better at image recognition than humans. But interpersonal skills like conveying health care information to a patient are not as easily or effectively performed by machines, he said.
“In almost every occupation, there are at least some tasks that could be affected, but there are also many tasks in every occupation that won’t. That said, some occupations do have relatively more tasks that are likely to be affected by machine learning” Brynjolfsson said, noting that a job like a concierge could be, and is being, mostly replaced by services based on machine learning from companies like Google. Jobs like massage therapists, which don’t have much potential for machine learning, are likely to be the least affected, according to the study.
The researchers recommended looking at the tasks within each occupation that have high potential to be automated by machine learning, separating them from the tasks that do not, and reorganizing the job to match those developments. Machine learning could be doing the tasks it is ideal for, they write, while human labor could be freed up to do more of the activities machine learning is not well-suited for, with a net effect of increased profitability.
That’s not to say new developments in machine learning couldn’t have a wider impact on jobs and the economy in the future. “Matching the evolving state of the art in ML in the future will require updating the rubric accordingly,” they write.
Artificial Intelligence and machine learning will soon replace trial and error as business strives to improve product performance. Rather than serving as a replacement for human intelligence, AI is seen as supporting tool. AI is kind of the second coming of software’s said Amir Husain founder and CEO of machine learning company Spark Cognition. It is a form of software that makes decision on its own and is able to act even in situation not foreseen by the programmers.
The world is changing and AI will certainly the part of future. Along with IOT, AI has the potential to dramatically remake the economy .Realizing the full potential of AI in future will require graduating from an automation view to a design view of the creation of hybrid human machine system.
With this machine learning training and courses are becoming interest and attraction among the students. Aedifico Tech is simply the best premium institute of the lot that provide an in depth knowledge about AI and machine learning along with hands on projects and exposure to live implementations of the concepts.
Get in touch with our mentors to know more about Machine learning and learning opportunities we have for you upcoming talented youth.