Why is data eating your culture before breakfast

A short, insightful read for busy executives.?

Is “the truth” the most important driver for you, or perhaps shareholder primacy, fairness, or, say, “equality and transparency?”? Your prioritised driver determines how you categorise and order information. This information structure, along with purpose and culture, determines what your business is optimising for and how you can deliver it. How you interpret the world according to your framing either creates conflicts or alignment for management.???

The same is true with data.? How you structure data determines your ability to fulfil your purpose and the culture you will have. Until now, you could ignore the topic of data philosophy and data ontology or delegate it to the CTO, but if #AI now drives your future thinking and strategy, that option of ignoring and ignorance is no longer available.

It is now imperative for businesses to ask and answer the question, “What is our data philosophy?”

There are a few major camps in mainstream philosophy with a wide variety of nuanced smaller ones - the same is true for data philosophies.? A personal alignment to one camp is dependent on an individual's stance on ontology.? Ontology is the top-level structure of categorising and organising information according to relationships. Your top-level worldview is quite specific to how you think, your culture, and your experience. Ontology is often an unspoken and unseen assumption, but its very existence frames how you and data perceive everything.? Your beliefs, morals, ethics, faith, hope, vision, motivation, incentives and actions are all informed by your ontology.??

The 17 UN sustainable development goals are different ways to organise information for action and highlight that there are conflicts everywhere. Should we feed everyone and remove hunger or reduce CO2 and make the world more sustainable? Everything involves a compromise (because we have limited and finite resources), and individuals and groups will naturally push towards their comfort zones (least stress, tension and past experience). That is how we reduce our own personal conflicts and tensions within ourselves—it is an alignment and an attachment and helps us make the wrong decisions! Some label it as bias.

In data science, there are three emerging camps of “data philosophy.” This has an important impact, as how you (and your company) organise data matters. It frames how your data sees your world, and therefore, it directly correlates to how and what recommendations reach the board, whether you can create value, and how you manage risk.?

Ontology (the structure of data and information) matters. It matters more today because of our increasing dependence on data and emerging ML/ LLM/ AI models. If you are new to this, this article will make much more sense if you read these two articles first: Data is Data and Data Ontologies. Data is Data unpacks why data is not what you already assume it is.

However, before the camps, “data is oil” was singularly the worst idea propagated. Those who repeat it (often without thinking but like it) tend to have a data philosophy by default, which needs to be unpacked.

Emerging data philosophies

1. We are only data. Data can and will be able to represent you in every detail. Your data will know better what you will do next than you do.? We (humans) are really very simple rules-based engines (chemistry) - but the scale and scope of interactions make the simple look complex. But fundamentally, we are only data.? This view was popularised in the mid-naughties by Eric Schmidt, the then-CEO of Google, who implied, “Google will know better what you will do next than you do.”? ?
2. We are far more than data. No matter how much data can be gathered about me, it will never be able to understand or represent me. The complexity of how my DNA interacts with my gut microbiome and brain, which depends on my nutrition, means I don’t know what I will do next. We are far more than data. This view has grown in popularity as we discover more about the actual scientific complexity of the human body/mind and what makes us think what we think in the context and environment in which we live.?
3. The agnostic. Data philosophy is rubbish (and probably finding reading this article).? Data is a tool. Data is oil, sunshine, and labour- it is just a thing to be exploited and manipulated by us.? We are in control.? Data has no meaning; wisdom is a label and is no more than more data.? Note: the idea that data will tell you what to do is not a philosophy or ontology but a capability.? You can always bend data to get it to give the direction you want.? The agnostic believes data will confirm what they want to do.? You are actually agnostic to the risks, consequences and impact - just like the extraction and use of oil.

In the interests of transparency and honesty, I have been in each camp over time, and I will continue to swap camps as I learn and discover more. My personal data philosophy is still very much in its infancy (after 20 years of thinking about this), but I am more aligned with one that I will happily explore with you if you ask.

There are many nuanced camps; here are three more that are set out to challenge you further about where you sit. These bridge or straddle the major camps above with subtle thinking that helps clarify why holding a pure and dogmatic view is dangerous. I prefer to think about the shades of colour rather than the black-and-white data that forces it to be this or that. There is obviously an opposite view for each one.?

The machine is authentic.

Machines can learn everything from data, and once they have been learned, like humans, their views, opinions, and responses are authentic.

Here is me - I have learnt empathy (data). I am a neuro-minority, and one thing of many I have had to learn is empathy. For many, empathy does not feel like a learned skill or capability, but something you were born with is natural.? Because I had to learn it, does it make my empathy less authentic than yours as you did not feel you had to learn it?? If you believe that other humans who have learnt empathy can be authentic, then why do you believe a machine or AI cannot learn empathy from a data set (responses) as well and can't also be authentic????

Data has agency. Data via an agent can act.

Agency is the ability to act, not the freedom to act.? Algorithms created from data already control machines that act; they have the agency to act within certain constraints and boundaries and can perform and improve.? Code is the same. Society and laws are the same - boundaries. Code and machines already have agency - what they don’t have (yet) is the freedom to choose to do something different (outside of their constraints).? At what point did you realise you could take the experience (data) from one field of expertise and apply it to another?? You were not granted the responsibility and needed no prior experience - just the ability to act (agency) to make a jump.? In your view, is that ability to jump called …. innovation, experience, invention, flexibility, resilience, adaptability, creativity, agency or freedom?? Has your view of machines being about to act too limited or too influenced by science fiction films?? Do you hold a view that has already been passed?? Data already has agency, but is that your philosophy??

Data can evolve independently

..... and it does not have to be consciousness.??

Advantage in the natural world is where mutation creates a feature that gives its host an advantage.? The living thing did not will or imagine the random mutation into existence; it happened because of a flaw or error in reproduction.? A beautiful example of this is moth (owl) eyes, where the moth did not will, create or think about them, but because they provide an advantage - they thrive with this error because the owl sees their own eyes, which fools the owl and gives enough time for the moth to escape.? Because data has errors and flaws, it raised a data philosophy question, “Do you believe errors (random mutation) can create advantage?”? It is not something you can “will” into existence; it just happens. This matters as it informs, affects and affects your views on data errors, bias and analysis.? In your thrust to remove bias - have you removed your ability to have an unseen advantage? (note - advantage is not competition)? Data is not conscious, but that does not mean that AI cannot evolve new traits and advantages.

Why is this important??

Understanding philosophical underpinnings can help shape how data is approached, interpreted, and applied in various domains.

Humans have studied humans, looking for patterns and, from this, crafted and refined theories (philosophies) about how we think/ behave for millennia.? You are probably utilising the modern ideals of behavioural economics as part of your design processes. Most management theories, including scientific ones, are built on philosophies. It is evident that all philosophies have a moment of truth, but none are always true.? So far, data gathered has been collected and analysed by other people. However, we have entered an age when the analysis and improvement of concepts, theories, and ontologies are created in real-time based on data using tools that are capable of self-learning at scale and speed humans cannot replicate.? The biases that create our paradigms and prevent us from seeing the obvious will be challenged. What is truth and creativity are examples of what is about to change.?

This is why you, as a leader or board executive, must create and refine a view of data philosophy and ontology for your company that aligns with your (personal and company) purpose and principles.? Because if you do not, your culture and purpose will continue to diverge from what your data is capable of enabling or delivering. You will not be in control of your strategy.? This is not easy, and it will definitely not be achieved in a one-hour side agenda on an away day; this demands deep reflection, consideration, guidance, and expertise.??

As a board, you cannot delegate such a crucial part of your future to the CTO, CIO, CIRO, or favourite consultancy to come back with a plan, as this is not the plan. This (your data philosophy) is how you will determine alignment between papers being presented for approval and purpose in the future —it is the very essence of governance and will be your legacy. It is the very construct of the culture and purpose of your organisation.?

Unless you are prepared - data will eat your culture before you even wake up, if it has not already.

data will eat your culture before you even wake up

The question you should ask at the next board meeting is “What is our data philosophy?”

The question you should ask at the next board meeting is “What is our data philosophy?”

Further reading

I have used a narrative and story above to make the point about different data philosophies; therefore, below is a more detailed examination of the major camps of data philosophy. These camps represent different approaches to understanding the nature, significance, and implications of data in contemporary society. Each is focused on a unique “aspect” of data, but they cross over and intersect—each informing the other in complex ways.??

Data Constructivism: Data constructivists argue that data is not inherently objective but is instead shaped by human interpretation, context, and social processes. They emphasise the constructed nature of data and highlight how cultural, political, and historical factors influence what data is collected, how it is analysed, and what conclusions are drawn from it.

Data Empiricism: This view holds that data should be the primary source of knowledge and that theories or models should be derived from and grounded in empirical data. Proponents argue that data is objective and can reveal insights that may be missed by pure theoretical reasoning.

Data Interpretivism: This view holds that data is inherently subjective and that its meaning is shaped by the social, cultural, and historical contexts in which it is generated and interpreted. Interpretivists argue that data cannot be separated from the values and biases of those who produce and analyse it.

Data Realism: This camp emphasises the objective existence of data as a reflection of reality. Proponents of data realism argue that data represents concrete aspects of the world and can be discovered or observed through measurement and observation. The focus is on the technical aspects of data collection, storage, and analysis.

Data Rationalism: This perspective emphasises the role of reason, logic, and theoretical frameworks in making sense of data. Rationalists argue that data alone is not sufficient and that it must be interpreted through the lens of conceptual models and prior knowledge.

Data Positivism: Positivists believe that data should be the sole source of factual knowledge and that subjective or metaphysical considerations should be excluded from scientific inquiry. They emphasise the importance of verifiability, replicability, and objectivity in data-driven research.

Data Pragmatism: Pragmatists focus on the practical consequences and utility of data rather than its objective truth or subjective meaning. They argue that the value of data lies in its ability to solve real-world problems and inform effective decision-making.

Different to Philosophies are skills/ capabilities?

Data Ethics: This camp focuses on the ethical implications of data collection, use, and dissemination. It explores questions such as privacy, consent, transparency, fairness, bias and accountability in the context of data-driven technologies and practices. Data ethicists seek to develop principles and frameworks for responsible data stewardship and governance.

Data Ontology: Data ontology examines the fundamental nature of data and its relationship to other entities in the world. This includes questions about the ontology of data objects, their properties, and the relationships between different types of data. Data ontologists may draw on philosophical traditions such as metaphysics and epistemology to explore these questions.

Critical Data Studies: Critical data studies adopt a critical perspective on the role of data in society, focusing on power dynamics, inequality, and social justice issues. This camp examines how data is used to reinforce or challenge existing power structures, shape public discourse, and perpetuate or mitigate social inequalities. Critical data scholars often draw on interdisciplinary perspectives from sociology, anthropology, and cultural studies.

Data Aesthetics: Data aesthetics explores the aesthetic dimensions of data visualisation, representation, and interpretation. This camp examines how data can be transformed into visual or auditory forms that evoke emotional or aesthetic responses and how aesthetic considerations influence the perception and interpretation of data.

New Risk.? Are you aware that combining different data sets with ontologies does not work the way you think it might?? We know you cannot compare or combine apples and polar bears, but often executives ask their staff to combine data from different sources, hoping to find the magic alexia. However, all they do is create quantum risk. Combining differing ontologies creates data and information about something that does not exist in either data set. Combining data sets that look similar but from different ontologies creates (new) risks you didn't know you had.?