Everyday AI could become as ubiquitous and necessary as electricity

Artificial intelligence (AI) is becoming ubiquitous. It provides directions while we drive, answers our questions, offers music recommendations and powers a growing number of business processes in the workplace. 

In fact, AI is working its way into so many aspects of our personal and professional lives that my company has begun to refer to it as “everyday AI.” Soon, I’d argue, it will become as ubiquitous — and necessary — as electricity.

Yet, despite the progress, we’ve only scratched the surface in the potential ways that AI can, and no doubt will, change business and the world. Gartner has forecast that it will take until 2025 for half of organizations worldwide to reach what Gartner’s AI maturity model describes as the “stabilization stage” of AI maturity or beyond.

So, there’s still a lot of work to do to put all the AI pieces into place — the software, systems, machine learning (ML) algorithms, data pipelines and governance controls. As more organizations build and expand this kind of AI infrastructure, the benefits of everyday AI — including productivity, efficiency and data-driven insights — will increasingly reap rewards for businesses of all sizes.   

AI: The electrical evolution of the 21st century

In many respects, AI is like electricity was in the 19th century: nascent, promising but untested, potentially dangerous without safety precautions, and with huge implications for the way it could potentially transform society. 

Some of the underlying forces of electricity — magnetism, polarity, electric charge — were understood well before Benjamin Franklin, in 1752, flew a kite with a metal key attached to demonstrate that lightning was nature’s form of electricity. But it wasn’t until the 19th century that the emerging field of electrical engineering put the necessary pieces into place to make electricity available to people everywhere, including one of the world’s first “direct current” power networks in New York City. 

Today, you could say we are in the early light bulb or electric-toaster phase of AI. It’s still proof of concept in many places, and the AI grid — the systems and software that extend AI applications across global businesses and to consumers — has yet to be fully implemented. 

And yet, many people already have their first experiences using AI, directly or indirectly. Amazon’s Alexa and Apple’s Siri, those omnipresent voice-controlled virtual assistants, are two of the most widely recognized examples, but there are many others. Website chatbots, auto-correcting text tools, and facial recognition for authentication are pervasive and easy to access.  

In these everyday situations, we understand that AI algorithms are driving business processes and user experiences. We know AI when we see it. 

But AI also works in less obvious ways. It’s often behind the scenes in loan approvals, supply chain optimization, and manufacturing automation, for example. 

Approach with caution

So, the line between when AI starts and stops can get fuzzy. I once made this point to a roomful of data scientists in an exercise I called “Do I Do AI?” The idea was to show that we do not always recognize when AI is powering processes such as auto-generated news articles, which are now commonplace in the media. 

At the time, I suggested we use four characteristics — learning, interaction, perception and goal seeking — to gauge the relative “AI-ness” of things on a case-by-case basis. This is the intelligence aspect of AI and increasingly, it will manifest in more and more of our digital experiences. 

But we need to be careful here. As more businesses move in this direction, it’s essential that we have the safety mechanisms in place to avoid the AI equivalent of an electric shock. One of the reasons the electric grid works as well as it does, occasional outages notwithstanding, is that circuit breakers and other safety apparatus are designed to avoid mishaps. 

AI needs its own grounding rods: Ways to minimize AI bias, deep fakes, privacy invasions and other unwanted consequences. I recently saw a report that more than a dozen autonomous vehicles got jumbled up at an intersection in San Francisco, resulting in a two-hour traffic jam. It’s all part of the AI learning curve, so we must be prepared. 

Trust and transparency in everyday AI

The key to long-term success will be establishing trust in AI through transparency. Not only must AI behave as expected, but we must be able to demonstrate and even prove that’s the case. This is why the concept of AI explainability, a methodology to track and validate AI-driven processes, is vital. 

How can business leaders ensure that their organizations have the appropriate AI governance framework in place to meet this high threshold? Fundamentally, governance requires prioritization and standardization of rules and processes in the design and deployment of AI. In addition, it’s important to align AI outcomes — not just with financial results but with non-financial objectives such as sociological and environmental goals. 

As AI advances on this path, new use cases will proliferate and drive increased adoption. Just as demand for electricity grew with the advent of light bulbs, toasters and coffee percolators, so will AI be accelerated by the emergence of virtual assistants, autonomous vehicles and smart homes.

In much of the world, electricity is now considered a necessity. There’s a corollary as AI progresses from pilot projects to enterprise-wide initiatives in business. The day may be coming when we can’t imagine our work and personal lives without the illuminating capabilities of AI. 

Florian Douetteau is CEO of Dataiku.