For a while now, I’ve noticed a troubling trend. The conversation around AI has become hyper-focused on individual performance. Everyone seems to be scrambling to use AI to become “better,” driven by a fear that they’ll either be replaced by AI or by someone who uses AI more effectively.
This narrative has completely pushed aside one of the most crucial topics of all: collaboration. Instead of cooperation, the focus has shifted entirely to competition. We’re losing sight of what truly drives innovation and transformation. It is and always has been collaboration.
Interestingly, if you look at the latest developments in AI itself, a different story emerges. The field is moving rapidly toward agentic AI and multi-agent systems. What developers have learned is that a single AI or a single large language model (LLM) can’t solve every complex problem on its own. It requires multiple AIs and different LLM, often with different specializations, strengths, and weaknesses, working together to achieve a goal.
Beyond AI: The Power of Human Intelligence
This is where we need to reframe our approach. A multi-agent system shouldn’t just be a collection of AIs. It should be a blend of artificial intelligence (AI) and human intelligence (HI). We need to intentionally model the collaboration between these two types of intelligences.
This means we must stop thinking only about the technology and start thinking about the organization. We need to consider:
- Who are the generalists and who are the specialists?
- What are the specific strengths and weaknesses of each human and each AI agent?
- How can we best organize these different intelligences to work together effectively?
Ultimately, the technical side is not the most decisive factor. The key is how we organize ourselves. This brings us back to the Datentreiber framework: Technology, Organization, and People (TOP). But now, the “P” for People needs to be expanded to include not only humans but also “artificial persons” as “P” for Partners—how we integrate AI into our teams and workflows.
A New Model for a Smarter Future
This is a powerful argument for what is called collective intelligence. It’s a call to action to rethink our hierarchical, siloed systems. We need to adopt the fluid, goal-oriented structures we see in cutting-edge multi-agent systems.
Look at the latest developments with models like GPT-5. We’re seeing a clear pattern: simply adding more compute power, more parameters, and more data doesn’t lead to a linear increase in intelligence.
This is predictable if you look at biology. The intelligence of a species doesn’t scale linearly; there are initial delays and later plateaus. We’re seeing the same thing with LLM now. We have to pour in massive amounts of resources for only marginal gains.
Why Collaboration Is Our Best Bet in the Age of AI
Here lies the strongest lesson from biology. The biggest jumps in intelligence in our species didn’t come from a single individual’s brain getting bigger. They came from collaboration. We are social beings who have built families, societies, nations, and global corporations. All of these are testaments to the power of collective intelligence.
True intelligence emerges when individual intelligent beings—human or machine—work together intelligently. This is what we need to prioritize: designing the organization and the processes for how we collaborate to become collectively smarter than any single machine or human could ever be.
Disclaimer: This article was a collaborative effort between human intelligence Martin Szugat and artificial intelligences Gemini and ChatGPT. 😉
