Artificial Intelligence – Part 1: Opportunities and Challenges

General artificial intelligence – the idea that a computer can simulate (or surpass) human performance in all situations – remains a distant goal, perhaps never to be achieved. But by breaking down human mental activity into discrete processes, and by optimising specialised processes for speed and accuracy, computers can excel far beyond human performance.

This highly specific form of “narrow AI” has already been adapted to countless fields – first for the performance of predictable and repetitive tasks, then for more complex activities such as speech recognition. Next came processes requiring deeper analytics, such as evaluating legal documents and sorting user data from online platforms. As the circle of AI competence widens, the area of human superiority will shrink; the UK government’s recent announcement of driverless cars by 2021 is but the latest example of accelerating computer achievement.

The trajectory of AI’s progression is fairly straightforward and, at this point, inevitable. But as a recent analysis by McKinsey Global Institute shows, the implications for business are extraordinary. The essential lessons to remember are as follows:

  1. Where AI is qualified to perform a task, it almost always represents an improvement. Traffic management, medical x-ray analysis, diagnostic testing, predictive online shopping recommendations tailored for individual users based on their personal data – these can be performed faster and often more accurately by computer than by a human specialist. Machines can make mistakes, particularly during the initial rollout of a new programme, but they learn quickly and soon outperform their carbon-based counterparts. Recent studies show AI offering significant improvement over traditional analytics techniques in nearly 70% of potential use cases.
  2. Because AI can outperform human talent, early adopters gain a cumulative advantage over latecomers. Companies with a focused AI strategy can use the technology’s benefits to find more suitable talent, optimise design and workflow, improve production methods, and train their employees to make the most of the new system. Each month that passes will see late adopters falling farther and farther behind their competitors.
  3. AI can do things we didn’t even realise were possible. Although Thomas Edison is best known as the creator of the light bulb, one of his other projects was at least as innovative: Edison invented the concept of an invention factory, where scientists were given freedom to run with their imaginations. A similar revolution is beginning to take place in sectors where research is driven by AI, as computers can be assigned to solve a problem without being constrained by the limits of the human imagination. Reports from many industries have shown that computers can come up with solutions that their human counterparts would never even think to try.
  4. Early adopters will get the first picks in a limited AI talent pool. Perhaps someday every public school will teach computer coding and robotic engineering; but at the present moment, experts trained to build or operate AI systems are being recruited faster than new ones are being created. Organisations that delay transitioning to an AI-enhanced business strategy may find that, by the time they are ready to hire, they will have to compete with a sea of other companies for a small pool of candidates.
  5. Indirect benefits from AI can be as important as its primary advantages. When an organisation sets up an AI enhancement for any given process, the employees previously responsible for the now-automated tasks can benefit from a reduced workload. This new-found freedom allows them to improve service elsewhere, through an increased focus on tasks that still require human abilities. AI can therefore improve the quality of output even in areas where it does not make a direct contribution, such as with soft skills in recruitment.
  6. The improved work environment created by AI can help organisations attract and retain quality talent. Rather than functioning as a replacement for a human workforce, AI can augment it – very similar to the way a smartphone can give its user new abilities. With employees relatively free from the need to perform monotonous tasks, they can appreciate the freedom they have to use their personal talents and creativity at work. People feel valued and appreciated when they are allowed to expand their skillsets in engaging ways, and interact more with their colleagues and customers. When the workforce can spend more of its time on satisfying tasks, morale tends to increase – as does employee retention.

And yet, despite these benefits, AI also brings its share of risks. The more a system is automated, the more it becomes vulnerable to malicious software and simple programming errors. Privacy issues also increase, even apart from the possibility of outright hacking. Companies that embrace AI should simultaneously embrace a new set of policies to protect themselves from these new dangers.

Moreover, if companies late to adopt AI will fall behind their competitors, then the same may hold true for entire countries. Similar rules will also apply to the international workforce, where those who are unqualified to work alongside computers may find themselves struggling to adapt to the new economy. Large-scale social disruption could very well accompany the digital disruption that has already begun to occur.

Individual organisations may decide that such large-scale issues are outside their area of concern. But as with any disruptive technology, consequences tend to be felt more widely than we first expect, and much depends on how carefully we usher it into our culture.

For example, we may also be entering a new era of misinformation, with consequences that are yet to be fully appreciated. ‘Fake news’ can now be created convincingly out of thin air, with seemingly solid photographic evidence to support it. Even video is no longer immune to such manipulation, and a bad actor could cause real damage to a company or a political system by crafting faked media that could then be made to go viral.

If we are going to truly advance as an economy and a society, then we must begin planning for the full impact of AI early – or our system will be unprepared for the jolts that come along with it.