No doubt about it, Artificial intelligence, or AI, is upon us and it’s here to stay. Getting past the buzzword and recognising that AI is much more than an overhyped trend but rather a swiftly developing and promising field that can redefine our world as we know it, are the first steps towards joining the revolution.
Amongst the exponential growth, unchecked myths and misconceptions about what AI can and can’t do still persist, fuelled by Hollywood tales about artificial intellect gone berserk, virtual relationships and partners getting deleted by a phantom force, and other futuristic visions of a dystopian society.
Common myths about AI
The list of unfounded myths about AI is growing along with the possibilities; it’s appropriate to refute a few of the most widely spread if we are to have a productive discourse on the subject.
Myth #1: AI will take over all human jobs
To paint a clear picture, the truth is that some jobs will indeed be made redundant by artificial intellect and that’s something we need to embrace as a natural result of advances in our technological capabilities, now occurring at unprecedented rates. Quite frankly, this shouldn’t come as a surprise and isn’t unlike the sizeable shift brought on by the first industrial revolution when machines also replaced the need for manual labour and demanded new skills of workers.
A 2017 Gartner report theorizes that although AI could realistically eliminate approximately 1.8 million jobs, it will also create 2.3 million different jobs, thus becoming a ‘’net job creator” as soon as 2020. Peter Sondergaard, head researcher at the company, predicts AI will augment workers’ abilities, helping us perform more smartly, efficiently and productively. It goes without saying that, to stay competitive and relevant, the workforce will need to get retrained and re-qualified in the new skills required to work along AI.
Myth #2: AI can mutate and take over the world
Among some of the most common AI-based fears sit the uncertainty around humanity’s ability to fully control machine-led intellect, when it’s left to develop on its own accord, or how our relationships with AI-based machines could evolve in non-beneficial to us ways.
Not long ago, AI-based messenger bots developed their own language in the course of a scientific experiment conducted by Facebook, which was perceived by outsiders as evidence that robotic intelligence is unpredictable and thus, dangerous. When it comes to human – robot relationships, some studies have shown that attachment to robots could affect battlefield outcomes in war zones, or have undesirable effects in care-taking settings where a human becomes physically and/or emotionally dependent on a machine that couldn’t reciprocate such feelings.
Beyond these more innocuous examples, what if AI were allowed to develop its own war strategy, deploying indestructible robots and causing the end of mankind as we know it? It is clear we must develop strategies for dealing with the less predictable, desirable and positive aspects of incorporating AI into various aspects of previously human-only settings.
The research of Professor Mary Ann Williams at UTS (University of Technology Sydney), one of the leading institutes on AI in the world, aims to ‘’design autonomous technologies that can learn to delight and adapt in novel situations as they collaborate with people to achieve shared goals.’’ Building AI algorithms in ways that prevent development routes that go beyond their original designations could be one way of curbing concerns about safety.
Myth #3: AI will be the end of many sectors as we know them
AI can and probably will augment and improve the operations of numerous sectors, making them more time- or cost-efficient, though rarely eliminating them altogether. According to Professor Toby Walsh at the University of New South Wales, the sectors that could benefit the most from the early implementations of AI include automotive, healthcare, entertainment, finance and banking, human resources, smart home management, logistics, and even sports (unbeatable teams of robots, anyone?).
When it comes to AI, it’s perhaps prudent to let bygones be bygones and focus on the possibilities to achieve things previously thought impossible, like prolonging human existence through digital bots once the physical body has given up. If we can conclude anything with a degree of certainty, it is that the future will be interesting, to say the least.
Today’s AI growth is made possible largely by advances in hardware such as GPU chips developed by companies like Nvidia and ARM, as well as specialised machine learning chips, like those being built by Intel. The latter are part of a family of chips called the Intel Nervana Neural Network Processor family, or NNP, and are designed to boost deep learning training times. Another Silicon Valley startup, Cerebras, is rumoured to be building a similar type of chips, albeit doing so in a very stealthy mode up to now.
Dealing with inevitable change
In a few years, we will be looking at AI as a step in the continuum of humanity: making life easier better and longer, reducing the need for humans to perform repetitive, mundane or brain-intensive tasks. At present AI is a promising concept that we are yet to practically grasp; it is up to the software developers and data scientists to make the most out of the progress made in hardware and machine-learning algorithms.
“The social aspect of social robotics is so important, it is about understanding how people interact with technology, particularly technology that we might deem as artificially intelligent, and how we can create new technologies that better understand humans and interact with a social intelligence,” says William Judge, robotics innovation manager at Commonwealth Bank's Sydney Innovation Lab (CBA).
In the meantime, CEOs, CTOs and CIOs are all advised to investigate how incorporating AI into their products or services can enhance their customer value proposition, innovation cycles, or manufacturing processes before the competition figures it out. As Daryl Plummer with Gartner notes, “Those who seek value from technology-based options must move faster as their digital business efforts move into high gear. Speed of change will require variability of skills and capabilities to address rising challenges."
Copywriter: Ina Danova