An Introduction to Hybrid Intelligence

Hybrid intelligence may inform the future of work for knowledge workers of all kinds. For now, however, it’s still work-in-progress, conceptually and operationally.

This is according to Jacob Sherson, the Founding Director of Center for Hybrid Intelligence in Aarhus, Denmark. The center hosts an interdisciplinary research team exploring digital interfaces that facilitate collaboration between humans and algorithms.


What is hybrid intelligence?

Hybrid intelligence is a way for humans and AI (Artificial Intelligence) to interact where the AI is not just enhancing human intelligence, but responding in real time to the preferences of a human user. This, then, optimizes both algorithmic creativity and human creativity.



The potential impact and power of hybrid intelligence could eventually be transformative for global problem solving and have applications across sectors.

“Hybrid Intelligence is not just taking current job functions, and then augmenting them with AI and making humans a little bit more efficient,” says Sherson. “Hybrid intelligence is about trying to see how we can reinvent the way in which humans and technology co-exist, and what they can do together. Philosophically, we are trying to combine the best of both worlds.

But we’re not there yet.


Hybrid intelligence is still just a point on the horizon

While there is considerable worry amongst knowledge workers about the potential for artificial intelligence and hybrid intelligence to disrupt their careers, the truth, as Sherson notes, is that most AI implementations aren’t successful in corporations. And hybrid intelligence is a step beyond that.

“Hybrid intelligence barely exists today. From the industrial or corporate point of view, there has been this enthusiasm and hype, and maybe a sense of urgency to jump into these AI-related projects to avoid disruption. But a lot of these implementations fail in the real world,” says Sherson. “So, while AI works fantastically in the lab, it really fails on a large scale in corporate applications. “

“In fact, 70% of all corporate AI projects generate little impact, and half of the senior managers in companies find it difficult to integrate AI into their processes and systems. That’s the problem we need to solve.”

There are several reasons why neither artificial intelligence nor hybrid intelligence have achieved full liftoff. These include a still-developing ethical code to underpin the creation of hybrid intelligence systems, and the lack of a common language between non-technical corporate audiences and the data scientists designing AI implementations in corporate settings.

But the biggest challenge might be the simplest.

We really haven’t solved the understanding of how humans think, the roles of intuition and common sense and creativity, and what it all means for optimizing human interactions with technology.

These are some of the problems the Center for Hybrid Intelligence is trying to solve.

The team at the center ‘investigates research topics covering human individual and collective behavior and machine optimisation, and presents humans with specific tasks that are difficult for a computer to solve. Then, they compare the human and algorithmic solutions to each other, and create synergies between humans and AI through optimized games that include direct interactions between the two.’

These are steps on the way to true hybrid intelligent systems.


What does hybrid intelligence mean for knowledge workers?

What often worries knowledge workers is that even if the frameworks to optimize and implement hybrid intelligence are still being developed, isn’t there going to be disruption for workers? Isn’t there going to be disruption for me?

Sherson doesn’t think so. At least, not yet.



“Organizational changes are much, much slower and more complicated in real, corporate settings. So, while 100% of knowledge jobs are eventually going to be affected in some way, only a small minority of them are going to be replaced over the medium term.”

Meanwhile, organizations all around the world – including Exfluency – are taking incremental steps towards an artificial intelligence-enabled future and some form of hybrid intelligence.


Stitch Fix: Hybrid intelligence in action

According to Sherson, one particularly successful example of hybrid intelligence in a commercial context is the company Stitch Fix.

“Stitch Fix is a company that specializes in sending custom packages of clothing home to their customers. Each package has five pieces. And then, customers can pick whichever ones they want and send the rest back. What Stitch Fix is doing extremely efficiently is combining a data-driven approach with a human approach. Relying solely on a data driven approach would suggest that the company should only send black underwear and black t-shirts to customers because those clothes are, on average, the most likely to be kept,” says Sherson.



“What they are doing instead is combining that data-driven approach with a small army of individual stylists that individually pick out things and that gradually learn to engage with their customers on a personal level. The customer gets the best of both worlds. What works on average and something that is personalized and specialized.”


The coming revolution of hybrid intelligence

Despite the challenging road ahead to the creation of truly hybrid intelligent systems, Sherson is still excited about their potential impact.

“I believe is there is revolutionary potential in the coming years for hybrid intelligence if we can create systems that are built on the strengths of humans, of artificial intelligence and that work in the real world.”