When you hear them for the first time, certain concepts make your mind react with a click straight away. All of a sudden, they put the exact words to something that you’ve been doing, feeling and defending for years.
That’s exactly what happened to me at the beginning of 2024. I was in Las Palmas de Gran Canaria. David Macías had invited me to a session about the state of art in AI. This was two years after the massive rolling out of models of language into general use, in other words, the appearance of ChatGPT.
It was there that I heard the technical term Human in the Loop (HITL) for the first time. A term which was already in existence since the major American companies started to get involved with machine learning, supervised learning and big data, in around 2010.
During those years, I was developing business in a Canarian technological company, a partner of Oracle and of SAP, two of those American giants. Microsoft was in its process of product restructuring (that is, it was momentarily out of the game), while Google and Amazon were hurtling along with innovation on their platforms. Software as a service was already a reality, and that hit those of us who offered a business model based on licence sales, from which, by the way, we had made a good deal of money.
Today I have no doubt that this industry change which moved from product to service was the one which brought the value of people back into the centre of the conversation: the talent of entrepreneurial software application creators, developers and salespeople.
During those years, the technological sector exploded and multiplied itself over and over, and new brands, an almost infinite number, were created. Spotify had already come onto the scene, and I was all of 4 years old.
From technology to co-intelligence
I have personally lived through the innovation process in the technology industry: since the time when applications were being built in local physical settings, servers, hardware, all governed by systems engineers and software developers. Great people who were often working from company basements. Literally.
In each and every one of these projects, regardless of the software or of the investment size, the lesson was always going to be the same one: not one single data architecture will work if people don’t put their trust in it.
My engineering colleagues always trusted these architectures, because they were mega professional specialists who designed technological solutions with both accuracy and affection. I trusted them, because I trusted my colleagues, but also because my business role “could see” the innovation that was being brought into the business processes. But business users didn’t always have that trust. Not because the product wasn’t a good one. In my opinion there was no trust for two reasons:
1. Because people didn’t understand a lot of the terminology being used (and which continues to be used…) and if you don’t understand a message, you can hardly interact,
2. Because in most cases they were not included in this project. It was a technological project. Business was involved at the analysis stage and at the testing stage, in other words, when the supplier had already been chosen, the design of the project defined, the contract signed and various project milestones already paid for.
3. Due to a lack of leadership. The transformation that comes with a company adopting (not just implementing) adopting an innovation is strategic. And strategy is always linked to leadership. Even if the innovation doesn’t impact the entire organization and only affects a part of it, it’s necessary for someone to guide the team toward the promised improvements.
Well, all three things happened… and still happen. And some others that I didn’t experience, but they might have happened to you ; )
Other projects were fine-tuned from start to finish. Many, fortunately, because those years in the Canary Islands shaped a technological innovation ecosystem that is now world-class: Inerza, Atos, Contactel, Desic, Intelequia, IdecNet, Noray, Inventia, Singular Factory, SD3, Atlantis Technology, Sistemas de Datos, Velorcios, Aida, and so many other companies that made it possible for hundreds more to be created. Today, most of them are companies that export talent and knowledge all over the world. Apologies if I’ve left out any of the pioneers; it’s been quite a few years…
The brand
In short, yes, the business model based on services has meant that knowledge of people can be considered the priority, and this far, it no longer matters where it is: if it is in IT, in business, in external personnel, in universities… it is in people, full stop.
I confess that my current view on the emergence of large language models (LLM) is sometimes contradictory. On the one hand, I believe we are experiencing an expansive moment that allows us to access benefits on another level, but on the other, I’m concerned about its depth… and many other things.
That’s why I wanted to make this term my brand. Human in the Loop has defined, since the beginnings of AI, a collaborative system where machines process data and human intervention defines the result by validating, correcting, personalizing, and making the final decision on what we want to create.