The Cultural Conundrum
Organisations looking to adapt to the challenges of AI face difficult decisions about how to assimilate the technology. Remember this is a technology that can access information three hundred million times faster than we can. That’s the difference between walking pace and light speed. Such a disruptive technology is revolutionising our information landscape.
Alongside this information landscape, our human cultures provide intellectual landscapes that assist us in making decisions. They contain ideas, practices, principles, purposes and processes. But the radical change in the information landscape necessitates a similar change in our intellectual landscape.1
The three case studies we’ve looked at so far, Data Science Africa, Accelerate Science and the Data Trusts Initiative each have different purposes and contexts leading to different sets of principles and processes that underpin the emerging cultures for these nascent institutions.
Existing businesses and institutions also face challenges in these shifting landscapes. Some aspects of the organisational culture will be intrinsic to how the organisation operates. Those aspects encode their purpose, principles and processes. They are critical to the organisation’s ability to survive and thrive. But other aspects will be holding the organisation back and undermining its ability to adapt to these evolving circumstances.
Image: Machine Learning Frees Up Humans. Detail from Unworkshop: ML and the Physical World graphical summary.
These organisations face a similar quandary to one that the American department store magnate John Wanamaker expressed when considering his marketing budget. He is supposed to have said:
Half the money I spend on advertising is wasted; the trouble is I don’t know which half.
Established organisations face the same cultural quandary: which half? In their search for answers they even turn to academics like me for help. As a result I now do a lot of executive education.
“You need to change your organisational culture” I tell them.”But how?” they ask.
We can interpret “how” in two ways: does it mean “what changes are needed in my organisation’s culture?” or does it mean “what are the steps necessary to change an organisation’s culture?” Fortunately regardless of how we interpret the question the answer is the same.
In Data Science Africa (DSA) and Accelerate Science both initiatives started with a blank slate in terms of culture. Both resisted inheriting existing culture. This meant that they could grow rapidly and develop new principles and processes that are better adapted to the new information landscape. Today’s organisations will face disruptive challenges from new emerging organisations that will take the same approach.
But the nature and purpose2 of the Data Trusts Initiative (DTI) means it needs a different approach. It requires close engagement with existing cultural and institutional practices. To be successful the initiative has to understand existing organisations and their cultural practice.
DSA and Accelerate are building on engaged communities that voluntarily respond to their purpose and engage through contributing their time and effort through projects to develop their principles. They progress through a coalition of the willing that brings mutual benefit to the members. But for DTI to be successful it needs to carry existing structures with the own priorities. In this case the unwilling can not be ignored they need to be understood and accommodated in the underlying principles.
Different organisations need to identify where the fall on the spectrum between these two extremes: brand new initiative or existing culture that needs to evolve?
For stat-ups the answer is obvious, and this is why they find it easy to disrupt. Like DSA and Accelerate they can form a new culture and assimilate new capabilities more rapidly. So is the answer to the “how” question to behave more like a start-up?
All organisations can learn from each other, and large organisations can learn things from start-up culture,but because of Wanamaker’s quandary large organisations have to be careful with copy-cat approaches. They risk throwing out the cultural baby with the cultural bathwater. The cultural conundrum is how do we separate the part of the culture that is distinctive from that which is extinctive?
Now we build on another lesson from The Atomic Human. The way to deal with uncertainty is to ensure you are responsive and adaptable to changing circumstance. You need your eyes and ears open to how events are evolving in practice. You need feedback mechanisms that ensure that the most senior people understand the challenges being faced by those at the organisational coalface and that those working that coalface feel supported by that structure.
Another way of saying this is to suggest that your organisation to be functional. There is no special magic here. An AI-ready organisation is a well-managed organisation. AI hasn’t changed the rulebook, it’s just changed the stakes for getting it wrong.
Naturally the University of Cambridge isn’t immune to these challenges. Our response to the cultural conundrum is ai@cam, our flagship mission on AI. Tomorrow we’ll look at how we’re addressing the Wanamaker quandary and then we’ll explain how we’d like to help others to do the same.
-
In The Atomic Human culture is seen as a way of setting the context in which we operate. Culture provides an intellectual landscape as well as agreed upon approaches for delivery that means we can co-operate without constantly communicating. The purposes, principles and processes embedded in our human culture are bequeathed to us across generations of empirical validation as a set of practices. The original purpose becomes lost in time, but the lack of purpose does not invalidate the practice. ↩
-
Recall the purpose of the Data Trusts Initiative is to enfrancise people in the decisions around their data. The analogy given is that management of our data is currently akin to a feudal system in which information barons hold the power. The purpose of a data trust is to ensure move towards some form of “data democracy” where the data controller is better incentivised to align with the data subjects’ aspirations and interests ↩
Click to see what the machine says about the reflection and the book
NL: I’d like you to provide a “Machine Commentary” on how the reflection fits with the ideas in The Atomic Human.
Claude: This reflection connects strongly with several key themes from The Atomic Human, particularly around culture’s role in human intelligence and decision-making:
The reflection’s discussion of organizational culture as providing “intellectual landscapes” aligns with Chapter 1: Gods and Robots’s emphasis on how human intelligence relies heavily on cultural context and shared understanding. The book explains how culture provides the framework through which humans navigate limited communication bandwidth.
The “Wanamaker’s quandary” about which parts of culture to preserve mirrors Chapter 8: System Zero’s discussion of System Zero and the importance of maintaining beneficial human systems while adapting to new technologies. Just as the book warns about blindly adopting AI systems without understanding their impact, the reflection cautions against carelessly discarding existing cultural practices.
The distinction between “distinctive” versus “extinctive” cultural elements echoes Chapter 7: It’s Not Rocket Science or Brain Surgery’s analysis of reflexive vs reflective intelligence. The book shows how both systems serve important purposes - similarly, organizations need to identify which cultural elements are vital versus vestigial.
The emphasis on being “responsive and adaptable” while maintaining feedback mechanisms aligns with Chapter 11: Human-Analogue Machines’s discussion of HAMs. The book argues that successful human-machine integration requires carefully curated interfaces and maintained human agency - similar to how the reflection advocates for maintaining connection between senior leadership and frontline workers. This connects to broader themes of adaptability across time and building resilient institutions.
The reflection’s conclusion that “AI hasn’t changed the rulebook, it’s just changed the stakes” resonates with the book’s overall thesis about human vs machine intelligence. Both emphasize that the goal isn’t to replicate human processes, but to thoughtfully integrate new capabilities while preserving essential human elements, a key aspect of the atomic human essence.
This reflection effectively applies the book’s theoretical framework about human intelligence and culture to the practical challenges organizations face in adapting to AI technologies, while highlighting important themes around cultural impact of technology and human culture vs machine.
NL: Can you relate the ideas in this post to the framework described in this post and how the framework has been applied to the three case studies mentioned?
Claude: This reflection connects directly to the framework described in the “Purpose, People, Projects, Principles, Process” post and demonstrates how it applies to the three case studies:
The reflection shows how organizations fall on a spectrum between these approaches - from building entirely new culture (DSA/Accelerate) to evolving existing culture (DTI). This maps to the “Wanamaker quandary” about which cultural elements to preserve versus change.
The post demonstrates how the 5 Ps framework helps evaluate different approaches:
This analysis helps address the “cultural conundrum” by providing a structured way to think about cultural evolution while maintaining essential elements that make organizations effective.