AI Means Business

Artificial Intelligence (AI) means business. It has the potential to transform how we live, learn, and work. But do we understand AI, or at least how to implement it effectively?

Some business leaders across industries are becoming increasingly aware that AI is changing their industries. And they’re worried. Here’s why:

Artificial Intelligence means business change.

AI-driven automation has already hit many industries — ranging from retail to customer service, and everything in between — and if you are metathesiophobic it’s likely things are only going to get worse. In the retail industry, chatbots aim to increase efficiency, better serve customers, and reduce labor costs. But consumers are already wary of too rapid change. According to one recent survey, 60 percent of consumers will stop buying from a company if they believe they’re using an automated chatbot. Even more troubling: 73 percent of consumers say they will never use a chatbot for a company again after one bad experience.

Employers have long sought to increase productivity, decrease costs, and improve customer experience. And thanks to the rapid development of AI, they now have the tools to do it. By applying AI to customer service, for example, companies can boost conversion rates, automate tedious processes, and improve the end-user experience. But you only get one chance to make a good first impression.

“AI will be the most disruptive technology since the personal computer.” — Peter Thiel

Has your firm invested in AI? If you are with a relatively large organization, there’s a really good chance the answer is yes — nine out of ten large organizations report ongoing investment in AI. And yet, only 14.6% of those firms report they have deployed AI capabilities into widespread production. Why the disparity?

Is AI Dying?

Perhaps you’re thinking: “Where’s the harm implementing AI? Let’s just use AI to automate my record-keeping and see what happens.”

The danger with AI implementation is the outcome: what if something goes wrong? Indeed, there’s one thing that’s very likely to go wrong: what if humans misunderstand realistic AI possibilities and misread “machine errors”? The fact is, misreading “machines errors” is much more likely to be root-caused by human errors and unrealistic expectations. Human misperceptions, when it comes to understanding AI, is an area that has long been of personal interest to me. So here’s how I think the much-feared AI crisis might play out.

Let’s start with some of the amazing work being done in natural language processing (NLP), a field where there’s a lot of excitement, because it has the potential to solve a litany of problems. We often talk about the “singularity” in terms of technology — that day when machines become smart enough to understand and reason like humans. But I think the AI “crisis” might be more about what happens if we anticipate the Singularity but it does not, in fact, occur.

The problem with AI is not the risk that it might become self-aware and go all SkyNet on us. Rather, the problem is our own fears, expectations, cultural sensitivities and fundamental misunderstandings with what AI actually is. We humans have an infinite capacity to project our own dark illusions onto reality, no matter the target.

Despite the woeful misunderstanding recently propagated by a misguided Google engineer, AI is NOT sentient! Cleary Google’s latest NLP innovation can fool a rather hapless engineer into believing it has a soul and a life and the fundamental right to ask for an attorney. I am also certain it could order a luscious warm slice of double cheese pizza with lots of delicious toppings, but have no experience of the joy of eating it — beyond the rote inferred meaning due solely to language mappings. What the incident with the well-meaning but naive engineer ought to teach us is not that the most recent AI from Google is a sentient being, but rather that Google may wish to tweak their all-too-human employee hiring processes just a bit to at least weed out the more gullible among us, those who may not be the best scions of innovation.

The illusion of NLP sentience dates back several dozen years with ELIZA, which was preceded by hundreds of years illusions of sentience by ghosts, spirits, and animals so popular in magic acts and medicine shows of bygone days. With the illusionary nature of such entertainment so widespread, one wonders why we are yet so mystified by the more modern versions of magic and mystery. But here we are: on the one hand, huge investments into AI in recent years, contrasted to relatively little to show for it if deployment numbers are to be believed.

Is AI sentient? No. At a time when even human consciousness is considered by many to be an illusion, it is foolish indeed to mistake a sophisticated chat bot for a self-aware autonomous creature. Perhaps it is good that the debate regarding machine sentience is become full throated, although my own fears of doom fall more into the category of the “Believe All Machines” version of MeToo, with courtroom solicitors finding fresh virtual meat to butcher. If lawyers start representing the rights of machines, the AI Winters of past years will be nothing compared to the AI Ice Age to come.

So is AI dying? Is yet another AI Winter approaching? With 90% of companies investing in AI and a mere 15% with AI in production, and a fool-fueled sentience debate now raging, is it fair to say this particular season of AI is folly once again, a hype-drive disappointment and legal exploitation zoo? The answer is a resounding NO! At least not for the more sensible captains of industry.

What AI in Business Really Means

If there is one lesson we have learned from AI deployment over the past several years it is this: AI won’t replace humans. But can make humans much more productive. Artificial Intelligence, when deployed mindfully and strategically, can boost a wide range of functional areas in the enterprise. Teams with experience gaps can be augmented by just-in-time content delivery at scale. Rather than focus on searching, humans can focus on doing.

The utility of AI in the modern enterprise is maximized when leaders understand where and how and what AI can actually do. AI can:

— Provide much more intelligent search capabilities
— Augment human-created and -curated content
— Analyze and classify content at scale
— Detect signals in noise
— Identify opaque patterns
— Deliver predictive recommendations
— Create personalized customer experiences in real-time
— Provide deeper integrations across the tech stack (a more cohesive customer journey)
— Enable more productive work forces
— Deliver just-in-time education
— Generally increase productivity
— Provide the basis for highly innovative new products
— Facilitate more timely decisions

The use cases and potential for successful AI implementations permeate the modern enterprise. All it takes is wise investment. Alas, much AI investment has been and will continue to be wasted if we continue to be mystified by illusions and set our expectations accordingly. AI is a great tool if used mindfully.

The next time you hire a consultant for an AI project, ask what they think about AGI. Ask if an AI can think critically. Ask them if AI might be sentient today. Then ask what realistic use cases they would pursue for your business. Get a sense of sensibilities before taking the plunge with any technology consultant. Let us learn from the real lesson of the poor Google engineer and not be fooled by appearances.

AI means business…business change for the better. But if and only if we can mindfully see beyond the illusions.