Collective Intelligence vs. AI: Why Human Collaboration Matters
What do Wikipedia, the Manhattan Project, and the discovery of the structure of DNA have in common?
They are all triumphs of collective intelligence - the synergistic problem-solving power of groups.
As we navigate an increasingly complex world, understanding and harnessing this power is becoming essential for progress, and even survival.
In the race between AI and human intelligence, many assume AI is destined to win.
The AI algorithms of social media have already stripped us of the ability to have a meaningful online dialogue as everyone is walled off into their own online bubbles.
Instead of letting the AI of social media divide us, we should be pulling together with intelligent and respectful dialogue. We should look to our collective intelligence.
There is however a fair amount of evidence that the most readily recognizable form of collective intelligence, brainstorming, does not work. Individuals working alone generate more ideas that are produced in a brainstorm (1).
This raises doubts.
Is there such a thing as collective intelligence?
The Concept of Collective Intelligence
Evidence for Collective Intelligence
When you look into the literature specifically on collective intelligence, you find papers in top-tier journals such as Science, Nature, PNAS.
Researchers demonstrated in a factor analysis published in Science (2) that there is a performance factor for a group's ability to solve a wide variety of tasks that is not explained by the sum of the IQ of the group's members or the maximal IQ.
Collective intelligence is correlated 38-44% with the degree of social sensitivity in the group.
Another feature that was negatively correlated was the number of people who participated in the discussion.
The number of women in the group is related to social sensitivity, as women score higher in tests of social sensitivity.
A subsequent paper published in the Proceedings of the National Academy of Science (PNAS) looked at this across multiple studies and found a similar effect (3).
When Google looked at this across their organization they were able to confirm that the groups that had the highest IQ were not the ones with the highest collective intelligence.
“After looking at 250 attributes of 180+ teams and conducting hundreds of interviews, Google’s analysts concluded that it’s not who’s on the team that matters most, but how they interact with each other.” (4)
What this means is that the effectiveness collective intelligence is sensitive to how groups interact.
Unstructured free for all brainstorms do not work but if you pay attention to how a group is interacting a group thinking together outperforms individuals.
In 2010, 33 Chilean miners were trapped 2,300 feet underground after a catastrophic mine collapse. Their rescue seemed impossible, but through an unprecedented global effort of collective intelligence, all 33 were brought to safety after 69 days.
In his book Hidden Potential (5), Adam Grant details the work of the Chilean mining disaster as a dramatic example of collective intelligence.
What was remarkable is that the solution that ended up working was one suggested by someone not in a leadership position. The leaders had the social sensitivity to give the idea of an outsider a try.
The principal factor for success was the openness to ideas.
The whole effort was infused with a culture of openness and trust because they were facing a seemingly impossible task and nobody was certain about what they should do.
This extraordinary story is evidence of the potential collective intelligence - a force that may be humanity's greatest asset in the age of AI.
Factors Influencing the Emergence of Collective Intelligence
Collective intelligence does not just happen.
One of the biggest mistakes is the assumption that proximity is all that you need for collective intelligence.
As outlined above how people interact is the most determinant factor.
This has been my experience in working with consortium projects.
A typical structure for a consortium project is that it is broken down into work packages or working groups.
75% of the time, when there is a problem within a consortium, it can be traced back to the fact that the involved work package either stopped working or the dynamic within the work package is not optimal.
After more than 15.000 hours of facilitating collaborative meetings with the goal of achieving collective intelligence, I have learned that there are five factor couplets that are key to collective intelligence:
Safety and trust: The best collaborations start out collaborating for the sake of collaborating, not to achieve individual goals of one of the collaborating partners.
This builds up trust. Safety and trust are also essential for effective meetings. Meetings are where collective intelligence happens, particularly when working between organizations and geographically dispersed.
This is why emotional intelligence helps to guide collective thinking in a group.
Direction and focus: Direction and focus make it easier for groups to be more autonomous in finding ways to solve a problem or to make progress towards a goal. This is where strategy comes in.
Strategy is a set of integrated choices that give those working within that strategy the boundaries within which to innovate. There is also a difference between strategy and what most people think of as strategic planning.
Vision and mission: You deploy collective intelligence not when a problem is simple or a goal is easy to reach, but when a problem is complex and goals are challenging.
Vision and mission help keep everyone motivated to solve the problem and to put the time and effort into building up the safety and trust needed.
Structure and iteration: Collective intelligence is best when there is an optimal mix of diverse perspectives and expertise. This involves thinking across disciplines and engaging with various types of stakeholders.
This type of thinking necessitates effective communication, and communication between disciplines and stakeholders is similar to trying to converse with individuals who speak a different language.
When unable to speak someone else’s language, one resorts to pointing at things. In the context of problem-solving or achieving goals, the equivalent of pointing at things is creating a prototype or a first iteration. It can be as straightforward as the draft of a manuscript.
Structuring efforts such as building knowledge frameworks are also a key strategy for increasing consortium project impact.
New ideas and opportunities: The energy in any meeting skyrockets when a new idea or opportunity arises. This is particularly true when the idea or opportunity builds upon what you have been discussing in an unexpected way.
Facilitating new opportunities is one of the slow strategies for moving fast in consortium projects.
All of these factors make it easier to think and work together.
Practical Suggestions for Fostering Collective Intelligence
Use the above five factors as a framework for facilitating interaction. Bringing everyone together is not enough, it leaves too much to chance.
A facilitator who is neutral and attentive to the need for safety and trust can do a lot for driving these factors.
The job of a facilitator is about promoting divergent and convergent thinking to promote collective intelligence.
Everyone knows about the divergent thinking part.
This is the traditional brainstorm. A facilitator going around the meeting and asking everyone for their thoughts is straightforward.
What is not straightforward is maintaining a structure and focus. A traditional brainstorm is meant to be open and freewheeling. That does not work.
A facilitator needs to promote openness and divergent thinking while constraining the discussion. This is a convergent process, but it is not exactly convergence, more on that a little bit later.
Then there is the importance of inspiration. Reminding everyone of why you are doing what you are doing with a clear strategic vision and mission inspires people to think harder.
Having an affected stakeholder tell his or her story does this very well.
However, there is a risk that the story telling inspires more divergent thinking at a point when you are ready to converge.
Convergent thinking is more difficult. Many who call themselves a facilitator shy away from this part of the process. They hide in “it’s not my role to be an expert, that is for you - the team that is working together.”
The problem with that sort of facilitation is that if everyone is thinking about their own ideas, converging them with those of others can lead to a lot of ‘um, ah’ moments.
A good facilitator listens to ideas, then combines them and presents them back to the group.
The ‘ums’ and ‘ahs’ turn into ‘oohs’ and ‘wows’ as people begin to see the connections. Done well, it gets really good when they start adding in their own connections.
Then you are collectively converging.
Facilitation leads to some sort of output, even if it is just an outline. Outputs are key, even if they are not so good at first.
Prototypes are the ultimate form of communication. They allow disparate experts and diverse stakeholders to see what each other means.
A prototype can be a draft or an outline of a paper, a worked out project plan, or even something physical.
Working converge ideas that arise in a discussion is a form of rapid prototyping and iterative design.
The point is that the best way to build something together with collective intelligence is to pull the threads together in a complete concept, but then be willing to change that concept completely if need be.
Why does collective intelligence matter?
Collective intelligence is how you work to realize big concepts that will make a real difference.
The life sciences and healthcare are full of examples of major shifts that are stuck in the ‘‘nice idea phase. This might be expected.
On average, it takes 17 years to go from research findings to clinical implementation (6).
Of course, on such a timescale, it's going to take 50 years or more to fully realize the potential of the concept of precision medicine. However, such long delays are not a given.
One the biggest contributors to long delays in realizing such concepts is the lack of efforts to leverage collective intelligence.
Working in silos is a formula for incremental advancement.
Often the reason given is that the bigger study or the more ambitious project is not feasible. So, it is better to stick to the tried and true.
Publish a paper with a small group of collaborators that will contribute to the field and one day it will help tip the scales to the point where we will have precision medicine.
When you engage in collective intelligence, the pace of achievement expands exponentially.
Let’s take an example from another field - astronomy.
There are billions and billions of galaxies, but it would be nearly impossible to think about classifying all those galaxies. It would be nearly impossible to think about classifying all those galaxies, there are billions and billions of galaxies.
You would be considered perfectly sane if you said its not worth the effort to try. It is not feasible.
Yet, with collective intelligence that is not true.
By engaging the broader public to help classify solar systems The Galaxy zoo project (7) researchers have been able to assemble a dataset of 552K galaxies.
While it is still a ways away from the 1 trillion estimated galaxies, you can see how this act of collective intelligence has allowed for an astronomical amount of research. Please excuse the obvious pun.
Whenever you feel uncertain or frustrated by complexity, turn to others.
Collective intelligence is also about doing the work to prepare the ground for others to use what you have produced.
Take data structuring for example.
When you take the time to capture and structure your data in a standard format, it is much easier to compare that dataset to the datasets of others.
One of the most powerful approaches that such structuring unlocks is the ability to validate findings in independent cohorts. Imagine if all precision medicine studies were ‘validatable’ in this manner.
At this point, would we have precision medicine beyond precision oncology? This consideration is doubly important when you consider that what regulators want to see to accept new approaches is data.
What is needed is proof that tailoring therapies based upon biomarkers or molecular profiling is beneficial over standard approaches.
Underpowered studies that are not comparable don’t make much of a difference.
Another important aspect of structuring data is that it enables people who look at the dataset later to understand the meaning of a particular variable or measurement.
It might not be easy to believe, but even standard measurements in a given field can be hard to compare due to small differences such as the units, the timepoints of measurement, etc.
By structuring your data, you make it something with which you can collaborate with others through time.
Conclusion
The fears that AI will enslave us are unfounded.
As many have explained, AI is a set of tools that make it easier to access information.
As an enabler, AI opens up opportunities to put more effort into striving for collective intelligence.
Instead of spending your time with rote activities, you can spend more time trying to solve more complex and difficult problems.
In this article, you have learned ways you can leverage collective intelligence such as:
Facilitating interaction
Being bold and pulling disparate threads of ideas together into a coherent concept.
Turn to others when you are uncertain and overwhelmed with complexity.
Prepare the ground for others to stand on your shoulders.
Be part of the solution to what seem like seemingly insurmountable challenges.
Embrace collective intelligence in your own work and join the movement to achieve major shifts in medicine and healthcare.
We are now at a time where there is great potential to transform the lives of patients with concepts like precision medicine, disease interception and engineered therapies.
Collective intelligence is how you can contribute to realizing that potential.
Unlock the Power to Revolutionize Medicine and Healthcare
Are you driven by the passion to push the boundaries of medical innovation?
Imagine harnessing the collective brilliance of like-minded experts to tackle the most pressing challenges in the field.
That's the transformative potential of consortium projects.
Take the first step towards reshaping the future of medicine and healthcare.
Let's schedule a consortium strategy call and explore how we can forge a dynamic consortium project tailored to your vision.
Seize this opportunity to be at the forefront of medical breakthroughs.
Click on the button below or send me an email: scottwagers@biosciconsulting.com
References:
Mullen, B., Johnson, C., & Salas, E. (1991). Productivity Loss in Brainstorming Groups: A Meta-Analytic Integration.Basic and Applied Social Psychology, 12(1), 3–23. h
Woolley, A. W., Chabris, C. F., Pentland, A., Hashmi, N., & Malone, T. W. (2010). Evidence for a collective intelligence factor in the performance of human groups. science, 330(6004), 686-688.
Riedl, C., Kim, Y. J., Gupta, P., Malone, T. W., & Woolley, A. W. (2021). Quantifying collective intelligence in human groups. Proceedings of the National Academy of Sciences, 118(21), e2005737118.
Google (re)discovers collective intelligence Apr 10, 2018 by Adam McKenty
https://collectiveintelligence.ca/google-rediscovers-collective-intelligence/
Grant, A. (2023). Hidden potential: the science of achieving greater things. Penguin.
Morris, Z. S., Wooding, S., & Grant, J. (2011). The answer is 17 years, what is the question: understanding time lags in translational research. Journal of the royal society of medicine, 104(12), 510-520.
Walmsley, M., Slijepcevic, I. V., Bowles, M., & Scaife, A. M. (2022). Towards galaxy foundation models with hybrid contrastive learning. arXiv preprint arXiv:2206.11927.
Collective intelligence is greater than the sum of individuals working together.
Other articles
