At a recent media event in Singapore, the CEO of Boston-based AI tools supplier DataRobot blamed the media for hyping up the potential of AI, and over-stating the technology’s impact on jobs.
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
After all, with headlines like “38% of US jobs to be lost to robots by 2030” and more recently, reports on researchers creating an AI system that teaches itself new languages, it is understandable that the general public would view AI in awe – and perhaps with some trepidation too.
To be sure, AI developments have come a long way. In the 1980s, AI researchers had a hard time gathering the data they need. They also spent a long time running and validating data models due to the limited computing capacity they had then. Today, compute and data availability are not longer holding back data scientists, leading to huge leaps in AI capabilities.
Yet, the AI we are seeing today aren’t anywhere near human cognitive and intellectual abilities despite advances in deep learning that mimic how the human brain learns and processes information. By the way, scientists haven’t fully figured out how the brain works, so I’d say we’re still far from achieving singularity – a state where machines become smarter than humans.
What then is the current state of AI good at? For now, it is most apt at performing what I’d call “one-track mind” tasks, like picking out faces in a photo, identifying songs playing in the background and recognising voice commands. These so-called narrow AI applications have the potential to improve human-computer interaction and perform tasks like quality control in a production line with a high level of accuracy.
Further out in the horizon is general AI that can potentially do the stuff that humans do, such as solving a broad range of problems with some common sense. While the likes of IBM Watson have made some headway, there are different Watson machines that cater to different purposes such as medical diagnoses and wealth management. There isn’t a single Watson with multiple intelligences.
That doesn’t mean workers can stand still, because machines will only get smarter and better at performing a growing list of repetitive tasks that once took armies of people to do. What sets humans apart from machines are our ingenuity, tenacity and our common resolve to overcome humanity’s biggest problems – these are the qualities that will serve us well as we enter the machine age.