The human in the loop, and why it’s a problem

Generative AI is sparking the democratization of AI and a workplace revolution. While AI is becoming ever more sophisticated, you can be certain that AI won’t replace humans. However, a human using AI might replace one who doesn’t. The “human in the loop” using AI as a productivity tool is going to have a huge advantage over those who don’t know how to use AI effectively. 

AI is generating unprecedented productivity                                                     

The current breakthrough of generative AI is a turning point. It is the dawn of the transformation of many sectors, making a wide range of human knowledge jobs 2 to 10 times more productive.

The current hype doesn’t have much to do with Artificial General Intelligence (AGI), or human-level intelligence. In my opinion, AGI needs something fundamentally different to become capable of reasoning and planning, and is probably more than a decade away. Nonetheless, the transformation of human knowledge jobs based on generative AI has begun.

We now only see the beginning of the omnipresence of generative AI, and these systems will keep struggling with being accurate and not being able to explain themselves. But they’re good enough for many use cases in which the human becomes more of an overseer than the person actually doing the work. Obvious sectors that will benefit in the short run are copywriting, software development, content marketing, legal, customer care, HR, gaming, educational content creation and any administrative context in which documents, written texts and other data sources play an important role. Which basically applies to any company or organization.

A tool, not just a toy

AI critics rightfully point out that generative AI will not replace the human reasoning that is so often essential. These second wave pattern matching AI systems can’t reason by definition. But that doesn’t mean it is “a toy and not a tool”, even though people will keep using them to do fun stuff like rap about elephants in trees or talk to Yoda from Star Wars.

People who quote Amara’s law (“We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run”) are right. Adoption will not happen overnight. On the adoption curve, we’re past the innovation phase now. The technology is mature (again, for human-in-the-loop use cases) but we still need to go from “early adopters” to “laggards also adopting”.

Some people are making the point that the only thing we know is that most sectors will thoroughly change, but we have no idea yet how. They are right. You only need to follow a few generative AI people on Twitter to feel outdated if you don’t read up for at least an hour a day.

The great imbalance of the AI economy

Despite popular belief, the default effect of AI is unlikely to be job losses. Instead, AI can be applied to every sector, including HR, legal, healthcare, customer care, and more, with generative AI potentially leading to an improvement in quality, features, personalization, and the creation of better content.

I’m obviously in Camp Optimism, though I do see some pitfalls. There is an active debate raging on the impact of AI on society and its implications for ethics and the environment, which I won’t go into here. I want to focus on what I believe will be one of the biggest problems concerning humans using AI, or “the human in the loop”.

AI will amplify human productivity and intelligence. But obviously only for the people that know how to use it. Learning to use AI will be essential for any professional career. I don’t think a new big inequality will be created by the gigacorps or 1 percenters removing most human jobs – there are many reasons why this just doesn’t make sense – but the inequality between people that were already highly educated and / or digitally savvy and now learn to use AI tools properly, and the ones who are not, could easily increase tenfold.

I think teaching more people to use AI only solves part of the problem. Although I firmly believe no jobs will be lost, it doesn’t mean the available people will be a match for the available jobs. The human in the loop needs to be an expert at what he or she does. A person who knows better than the AI and can check all output, can add extra value. But mediocrity or inexperience will become a bigger problem, since the AI might just do better. It might seem counterintuitive (“Isn’t it an equalizer?”) but AI actually benefits the most talented and experienced people the most.

Finding the answer is our next challenge

It’s not that I’m pessimistic about this. We humans are remarkably good at dealing with things like this. Maybe new jobs can be invented that further improve lives? There are probably policies and ideas we haven’t thought of yet that can make this situation manageable for everyone. But the reason why I feel it’s so important to focus on this imbalance is that it’s already happening, and we don’t have the answers yet. I will never think of a junior software developer the same way anymore when we are hiring. We need to start building a proper mindset around this, before too many humans are left out of the loop.

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top