More human than human? The rise of business AI

Would you trust an AI to offer you a job? A Russian company is offering AI software to interview and recommend candidates for employment, at a much faster and more efficient rate than humans could ever hope to match. ‘Vera’ can perform all aspects of the recruiter’s job and is currently used by hundreds of Russian firms to assist in their hiring endeavours. She is a sign of the times; the use of automation and artificial intelligence in business is rising: but is this to be welcomed and what are the consequences for the sectors?

The automation of technology has long been a factor in the evolution of businesses: hardly a week goes by for instance without the latest development in self-driving cars, or a retail giant’s plan to deliver using drones. But it is in Artificial Intelligence that the biggest steps are being taken. Not just in terms of software for a single purpose, such as Vera, but for software that exists across the whole lifecycle, and breadth, of a business: AI used to streamline existing processes, plan new business ventures and enhance the supply-chain and management: essentially, the development of a business strategy with AI assistance at the core.

It is this ‘intelligent’ aspect of the machine that is what sets AI apart from simple rote machines: able to do more than just a pre-programmed task (Vera for instance, must be able to judge what constitutes a qualified candidate from a field where submissions may vary wildly, and be able to customise her written and oral responses each time), and learn as they go so as to become more efficient and useful to the business. Through extended machine learning, AIs are aimed to be able to carry out problem-solving and logical reasoning in the manner of human cognition. In the right situation – be it pattern recognition, prediction-making or autonomous searching – ‘the machine’ can be left alone to do substantial work (in terms of time and effort) without having to be explicitly programmed to follow narrow instructions.

The positive consequences of this could be unparalleled. If not only everyone in the business, but the system which underpins the business itself, is geared towards constant improvement then this would be a major shift. If it automates without losing the unique or creative aspects, then productivity and profits would doubtless go up.

The problem is what happens when AI is not enough. It may be able to learn and develop and thus expand its parameters, but the crucial angle of genuinely thinking for itself (which would turn it into a Virtual Intelligence) is out of reach, and as a result an AI will always be limited. Vera would not be able to react emotionally to a candidate and make a ‘gut judgement’ based on more than the strict responses she gets, and an AI underpinning a business system would not know how to react to an ‘edge problem’ that goes beyond the limits of its parameters.

This is particularly pertinent when it comes to intelligence. Already there is some degree of automation in intelligence-gathering: monitoring the dark web for desired keywords and analytics attempting to predict future events based upon an intelligence foundation. But what this cannot do is add the crucial human element that distinguishes artificial intelligence from true human intelligence: the ability to make seemingly illogical leaps of reasoning based on instinct or perceptions or to think ‘outside the box’ – something which is at the very core of analysis and intelligence work.

It is vital that this edge is not lost in the rush to adopt new and better technologies, as automation does not automatically equate with improvement and some businesses still are built around the human touch. This is not a business that can thrive on full-scale adoption; rather augmentation.

This is not to refute that AI has it uses, but it has to be appropriate to the needs and methodologies of the business it serves, rather than the other way around. Meanwhile, as for Vera, she continues to be used in over 100 Russian companies – an interviewer that never sleeps and never stops. It’s too bad she won’t live. But then again, who does?

Click here to download a copy of the article

Report Date : Aug 2018