Ask ten people what effect AI will have on the job market and you’ll probably receive ten different answers. It’s an emotive question that has devote followers at either end of the spectrum.
Some paint a picture of a dystopian future where robots have taken over the world turning humans into a sluggish, overweight race. But history shows that this vision is far more suited to science fiction, than it is reality. While AI in business is certainly fuelling massive changes, it’s a pessimistic view to say that it will replace humans.
In the widely noted 2014 study, The Future of Employment: How Susceptible are Jobs to Computerization?authors Carl Benedikt Frey and Michael A. Osborne state: “According to our estimate, 47 percent of total US employment is in the high-risk category, meaning that associated occupations are potentially automatable over some unspecified number of years, perhaps a decade or two.”
It’s a sobering figure, but as the report itself notes, this is only one aspect of the impact of artificial intelligence on employment.
Warnings of technology being a harbinger of death for the job market is nothing new. MIT Economist David Autor in Why Are There Still So Many Jobs? The History and Future of Workplace Automation notes that the Luddite movement of the early 19th century was one of the earliest examples, in which a group of English textile artisans protested against the automation of textile production by seeking to destroy some of the machines.
But in fact, that wasn’t the case, and basic economics intervened. Automation made it cheaper to produce fabric, which in turn led to more customers, which drove demand for more product. The job might have changed, but during the industrial revolution there was no shortage of work for semi-skilled labor.
James Bessen, an economist at Boston University School of Law looks at more modern examples in his blog post Automation Paradox. Software, for instance, made it cheaper and faster to trawl through legal documents; so law firms searched more documents and judges allowed more and more-expansive discovery requests. Likewise, ATMs made it cheaper to operate bank branches, so banks dramatically increased their number of offices.
These examples add weight to the point of view that, in all likelihood, rather than AI taking jobs, humans and AI will work in unison.
While many people warn that this time is different, that jobs are being sacrificed to AI in a much shorter timescale than with previous industry changing events, so far the figures don’t add-up. Rather than wiping out jobs, AI in the workplace is actually increasing the skill sets of workers, and therefore remuneration, across a wide range of industries from healthcare to clerical.
Artificial Intelligence is also improving workforce conditions. According to a report in the Economist, AI will help remove unconscious and conscious biases in the hiring and renumeration of staff. It also points out that AI will benefit employees in other ways such as ensuring the appropriate safety gear is being worn using intelligent scanning technology.
In addition, chatbots are being used by HR to support training activities too. This follows on from the success many chatbots have had as in-house advisors to call center agents in situations where a high turnover of staff can often impact on the consistency of answers and the knowledge to answer queries quickly.
The growth in AI is also opening up new opportunities in other areas of emerging technology closely linked with it such as Augmented Reality. Who could have predicted that Pokemon Go would lead the way to job creation? Already consumers are benefiting from the interest the game created. While currently that might only be seeing what a sofa online would look like their living room; it’s an industry expected to be worth $66.68 billion by 2022 as its use increases.
In 2016 Gartner suggested that IT leaders look for unanticipated consequences of the rise of IoT, saying that the secondary effects will be more disruptive than the initial digital change. A recent global survey conducted by Accenture cites that 61% of business leaders expect the share of roles requiring collaboration with Al to increase in the next three years. 54% placed human-machine collaboration as important to achieving their strategic priorities.
One area of business in which AI is gaining increased traction is in customer service, where enterprises have started to deploy artificially intelligent virtual customer assistants (often referred to as chatbots). Whilst this trend is still in the early stages, with only 4% of enterprises having deployed conversational interfacesaccording to a recent Gartner survey, 38% are planning to or actively experimenting in this market which is set for significant growth.
But how emotionally savvy are these chatbots? It might, if your chatbot uses conversational AI, be able to recognize sentiment. For example, it would be able to detect that a customer is angry and sarcastic because they are annoyed you didn’t deliver and be able to respond with the appropriate terms of empathy. Or it may be able to distinguish that “I want to go somewhere nice” is positive, vs “I want to go to Nice” is neutral, hence respond in a meaningful way. But it doesn’t replace the need for human connection, just as talking on social media doesn’t fulfil the same need as sitting down with a friend for a cup of coffee and a chat.
What AI does do however, is free up contact centre staff to deal with the emotionally charged issues. The one-off circumstances that your chatbot has yet to expect and so hasn’t learnt or been trained to give an appropriate response. Those situations where a real person can use their life experiences and combine them with your policies and procedures to arrive at a satisfactory outcome.
So, that bottom line is AI has its place; as do humans.
For example, machines are good at making sense of enormous amounts of data, of learning correct responses and statistically guessing the appropriate response. They are amazingly fast at processing; at making logical choices based on statistical rules. But when your customer’s expectations and satisfaction rests on a little empathy, wouldn’t it be great to be able to detect a shift in sentiment and hand off to a live agent – with all the appropriate background of the specific problem so it doesn’t need repeating.
And that is the crux of the debate – machines are all about data and humans are all about emotions. The decision to purchase with a particular company is more often rooted in emotional need than rational choice.
Undoubtedly, AI will replace humans in some roles such as process orientated tasks where RPA technology excels. However, computers are tools, not rivals. In every situation where technology threatens jobs, new positions arise, often because of changes brought about by technology. There will always be jobs that only humans can do, including designing, updating and enhancing the artificial intelligence technology itself.
While there is no doubt that achievements made in the field of deep learning or neural networks are impressive, it is not the fastest, nor the most cost-effective way forward for the average enterprise to develop conversational AI applications. Just like a child learning a language, an artificial system for natural language understanding needs human supervision. Even a statistical algorithm that learns from data can only do so from structured training data carefully curated by humans.
So why is there so much hype around algorithms? Perhaps because statistical algorithms are supremely useful for some purposes, such as aiding and guiding analysis of big collections of language data. And for some applications, neural network algorithms deliver very impressive results. Such algorithms have vastly improved speech recognition systems, the technology for mapping sound waves to text characters, which is the first step in processing speech.
But what seems like effortless communication to humans, poses multiple obstacles to a statistical algorithm. Unless training data are supplied in copious quantities, the signal—the meaning at the heart of the conversation—is lost in the statistical noise.
Put simply, when the algorithm is faced with too many ambiguities, too many options, and too little data, it gets confused.
The truth of it all is no matter how hard some may try to convince us otherwise, AI will not replace human emotional intelligence. Or, at least, not anytime soon. As an article in Inc put it recently, there isn’t a mainstream consumer machine that’s close to achieving full sentience.
Are there threats from AI that will affect all of us? Undoubtedly. The Inquirer recently reported Jim Al-Khalili as saying that without concerted action by the government, industry and academia, AI could end up “uncontrolled and unregulated”, with development monopolised by just a few powerful companies.
If enterprises sleepwalk into handing over their data to FAMGA on a silver platter, this could well be the case. The Cambridge Analytica debacle has already highlighted the value of personal data, even innocuous statements, and the need to protect it. It is one of the reasons we stress the importance of data ownership in Conversational AI.
Without data ownership, not only do enterprise lose valuable data insight, but it is also makes it harder to protect and secure the information. It’s clear that as data becomes the driving force behind businesses that data protection regulation around the world is going to increase. GDPR and China’s Cybersecurity law is only the beginning. Data privacy will become a major issue for everyone as the use of AI increases. Starting to address this now will place enterprises in a better position in the future.
Computerized automation does potentially put low skilled workers, whose jobs could be easily automated, at risk. Conversely, this may be a short-term effect while the labor market re-adjusts. As Carl Benedikt Frey and Michael A. Osborne say: findings thus imply that as technology races ahead, low-skill workers will reallocate to tasks that are non-susceptible to computerization – i.e., tasks requiring creative and social intelligence. For workers to win the race, however, they will have to acquire creative and social skills.
One of the auxiliary effects of the industrial revolution was how it changed education.
By the 1830s the British government started to fund education through charitable organizations. The newly introduced Factory Act meant that children working in factories attended school for at least 2 hours every day. By early 1900s civic or “red brick” universities were introduced to deliver vocational training covering areas from medicine and science to mechanics and engineering. Indeed, such was the change brought on by the industrial revolution that other countries following suit found that their success rate in capitalizing on the opportunity correlated directly to the standard of education within the region.
In a recent EY survey, 80% of respondents said that the greatest inhibitors to the uptake of AI in their enterprise was a lack of talent with the requisite skill sets. Perhaps the answer to the impact of AI on the workplace of the future is not to give dire warnings, but to look how re-educating and re-skilling workers will develop technology to take us beyond AI to the next big revolution.
Originally posted here