There’s no doubt that AI is all the rage! The stock market thinks so. The media thinks so. Business thinks so. The public sector thinks so. And I’ll bet even your UBER driver thinks so. There’s nothing wrong with all this buzz. It deserves the attention, because AI has made massive strides, recently in particular. But we really should take a breath for a moment. That’s because this resurgence of a field of study that has been around since the 1950’s has largely been fueled by machine learning. Machine learning has undoubtedly earned the spotlight—it’s getting results. Large amounts of data along with powerful advances in reinforcement learning, deep learning and causal inference are coming together to provide powerful impact in practice.

So why the pause? Honestly, I mean only for a moment. The reason is that machine learning puts data at the forefront, and it uses data to ascertain relationships between variables. This is how it forms its view of the world. Humans can be involved, as is the case in supervised machine learning, but overall the belief is that algorithms that fit to the data will show us the way. This is in stark contrast to the approach of imparting some structure on our models, based on decades of scientific knowledge we have about human intelligence, from behavioral science, and building out from there to test on data.

AI has always been concerned about the relationship of machines and humans, and this has reflected well in the research on AI alignment, human-centric AI, Humanistic AI and recently Causal Economic Machine Learning (CEML). These approaches put humans and human activities at the core of the discussion on AI. But much of the current discussion on AI puts the machine at the centre and speaks of stages of ‘evolution’ from Narrow AI to Human-Level AI (HLAI) and eventually to Super General Intelligence (SGI). Human intelligence (HI) is only invoked as a ‘benchmark’ reference.

History has shown us that we can’t ignore revolutionary technology, nor do we want to. Imagine a world where scientific knowledge, the printing press, industrialization or then Internet were totally suppressed? All massive breakthroughs come with great risks, but they also present the greatest opportunities to humanity. Most of us suspect that history will likely look back on Artificial intelligence as a true technology revolution.

With all this change and momentum, it’s now more important than ever to step back and think about this, and ask “What is our Goal?” Do we really want to create synthetic entities that function as a ‘species’ on their own, or do we want to maintain our hard-earned spot as humans at the top of the food chain, domesticating AI to always work for us? If we’re not explicit and persistent about this, the competitive human drive to accomplish more will push many to seek SGI in their thirst for accomplishment.

There is Middle Ground—Human AI

So, how do we get the balance right? Table 1 below shows the intelligence spectrum from AI (artificial) to HI (human) and some of the major categories that fit within it. In this brief article we won’t dive into all the elements of AI and HI at the ends of the spectrum. We’ll instead highlight the middle ground, where I believe we achieve the best of both worlds—Human AI (HAI).

Table 1

So Can People and Machines Really Get Along?

So back to our original question… can people and machines REALLY get along? I don’t have all the answers. And with revolutionary technologies like this, no one can make a credible prediction. But I do know enough to suggest that the answer is up to us—right now at this point in time. The role of each of us is to demand that AI serves humans and not the other way around and to support that mission through our actions. I have incredible faith in humans. We created AI and I think we’ll always have a leg up somehow. But, maybe that’s just me being human! It’s no doubt time for more discussion than ever, so share your comments, questions and concerns and let’s keep the dialogue going.