AI reality check

For the last three years, I’ve been active in the middle of the ‘AI storm’: I quit my well-paid job at the largest high-tech company in the south of the Netherlands, started an AI consultancy company, supported setting up several AI initiatives and had many discussions on the topic. Those discussions were at the extremes of “We’re too late as all AI innovation happens in the US and China” up to “We need to make all people in organizations data and AI aware including the receptionists.” Perhaps we should get a little grip on reality.

To stimulate people doing an AI reality check, I often ask them the following: if all companies are desperately looking for AI expertise, why isn’t there a large queue in front of my AI company door? Even if I would do a lousy advertisement job, with the ‘AI hype’ pictured in the media and lack of AI talent, there should be a queue.

Currently, only a few companies are really actively adopting AI technology. They’re motivated by the fact that in fields such as computer vision, planning & control and natural language processing, AI algorithms outperform traditional techniques. This creates innovation opportunities. These first movers are aiming at a competitive advantage at the cost of a risky and high investment. But the fact is, most companies haven’t yet made a move at all towards adopting AI.

Five years ago, nobody talked about AI. Currently, everybody seems to be an expert. There’s a lot of talk but no content – that is, real-world demonstrations of AI applications. Before more companies start to move, more AI content needs to be available showing the business benefits. But a lot of the buzz is created by parties that often lack the content but want to sell the potential. The usual suspects here are consultancy companies, startups, educational and research institutes and more. While there’s good AI content to be found among these parties, companies can generally not judge by themselves what’s good and what isn’t, as they lack AI expertise.

The answer here is education. But with the current AI education offering, companies face the same problem: how to judge what education is good and what isn’t? The offering is so wild that we even have national AI courses for everybody, including the receptionist who needs to become data aware – whatever that means for this person anyway. I understand that the introduction of a new technology is an exciting moment. Even my parents, who are in their seventies, are interested in watching an AI special on TV to catch up. But asking them to follow a course on AI is useless. They would be helped more when their smart speaker with ‘AI inside’ would understand more of what they’re telling it.

When I started my AI company, I wanted to share my 25+ years of AI experience with companies by training their employees to become AI experts themselves. While providing the training over the last few years, I realized that this perhaps isn’t the best, or only way, to support the industry. You can’t train an AI expert in a few days, nor a few weeks. AI is a profession. I’ve therefore extended the strategy of my company to support industrial AI adoption by developing products with AI inside that non-AI experts can use. No need to be an AI expert yourself.

Should your organization act on the AI storm? Yes, of course! AI technology outperforms traditional techniques in many areas, impacting many businesses. Should everybody be an AI expert? No! Organizations use many other technologies without everybody being an expert. In the coming years, more and more AI applications will be demonstrated and the business benefits of AI will become clearer. This will drive demand for AI technology. A large part will be served by products with AI inside, provided by innovative companies, and some part will be developed by your own AI experts.

So, let’s stay calm but do move.