History of progress: from Neolithic to Artificial Intelligence
Humans, their brains and ways of thinking

Messianism in AI times
How Scientific- and System-Thinking Can Get You a Competitive Advantage


Cedric Berger, PhD MBA (https://www.linkedin.com/in/cedricberg/)
July 11th, 2024
SWARM - The Human + AI Symposium

Profit

  • Who are "we"?
  • Is there a single "we"
  • Are we ready to make compromises?
Medinorma.003.png
‎DIKW variations.‎1001.png
Pasted image 20240708172402.png
Pasted image 20240708171955.png
‎Humans&AI.‎022.png

Minimise pain
Maximise pleasure

Profit

Pasted image 20240708172222.png
Pasted image 20240704093139.png

Human beings: my life will be easier, more comfortable, less painful

Business: new opportunity to make money

Nature biotech-Economy of promises.png
‎Humans&AI.‎023.png
‎Humans&AI.‎021.png

Hope

Not because of Neural Networks (these kind of ML algorithms are available since the 80's)
Because of:

  • hardware GPUs used for NN
  • significant up-front investment to train a LLM on internet
  • offering a service
Pasted image 20231025090808.png
Pasted image 20240625132333.png
2021-MAD-Landscape-V2-Small.jpg
‎Humans&AI.‎020.png
Big data landscape in 2012.png
Pasted image 20240625132544.png
The business of digital transformation.005.png
‎DIKW-Digital twining-KGs.‎014.png
Pasted image 20240404072405.png
Digital twining.006.png

They cost of lot of

  • money
  • energy
  • raw material

-> not sustainable

MAD_landscape_2023.png
Pasted image 20240625132733.png

They are not reliable (yet)

  • hallucination
  • biases
  • unspecific
  • does not understand the world but probabilities of words association from white English-speaking males

Their precipitated frenetic usage is due to

  • humans hoping to reduce pain creating artefacts
  • liberalism always seeking for increased productivity
  • greed for profit (economy of promises)

System thinking

Systems thinking is a way of making sense of the complexity of the world by looking at it in terms of wholes and relationships rather than by splitting it down into its parts.

It has been used as a way of exploring and developing effective action in complex contexts, enabling systems change. Systems thinking draws on and contributes to systems theory and the system sciences.

Wikipedia

System thinking.png

Pasted image 20240506150213.png

  1. Dynamic thinking
  2. System-as-a-cause thinking
  3. Forest thinking
  4. Operational thinking
  5. Closed-loop thinking
  6. Quantitative thinking
  7. Scientific thinking

From ChatGPT-4o

Systems thinking is an approach to problem-solving that views "problems" as part of an overall system, rather than in isolation. It emphasizes understanding the relationships and interdependencies between the components of a system to gain insights into its behavior and dynamics. Systems thinking is particularly useful for addressing complex and multifaceted issues where traditional linear thinking might fall short.

Key Concepts in Systems Thinking

  1. Interconnectedness: Recognising that elements within a system are interrelated and that changes in one part of the system can affect other parts.
  2. Feedback loops: Understanding that systems often have feedback loops, where outputs of a system are fed back as inputs, either reinforcing (positive feedback) or balancing (negative feedback) the system.
  3. Emergent properties: Acknowledging that the whole system can exhibit properties and behaviours that are not evident from the individual parts alone.
  4. Boundaries: Defining the limits of the system being studied, which helps in focusing the analysis and understanding the context.
  5. Holistic perspective: Viewing the system as a whole rather than focusing solely on its individual components.

Steps in Systems Thinking

  1. Define the problem: Clearly identify the issue or problem you are trying to understand or solve. Ensure that you consider the wider context and not just the immediate symptoms.
  2. Identify the system: Determine the boundaries of the system related to the problem. Identify the key components, stakeholders, and their inter-relationships.
  3. Map the system: Create a visual representation of the system, such as a causal loop diagram or a systems map. This helps to illustrate how different elements are connected and how they influence one another.
  4. Understand feedback loops: Identify and analyse the feedback loops within the system. Determine whether they are reinforcing (amplifying changes) or balancing (stabilizing the system).
  5. Identify leverage points: Determine the points within the system where small changes can lead to significant impacts. These are often the most effective places to intervene in a system.
  6. Develop hypotheses and test them: Formulate hypotheses about how changes in certain parts of the system will affect the whole. Use simulations, models, or small-scale experiments to test these hypotheses.
  7. Implement solutions: Based on the insights gained, develop and implement strategies to address the problem. Ensure that these strategies consider the system as a whole and the potential for unintended consequences.
  8. Monitor and adapt: Continuously monitor the system to see how it responds to the changes. Be prepared to adapt your strategies based on new insights and feedback.

Example of Systems Thinking in Practice

Consider a company experiencing high employee turnover. A traditional approach might focus on increasing salaries or improving working conditions. However, a systems thinking approach would map out the entire system, including factors like company culture, management practices, workload, career development opportunities, and external job market conditions. By understanding the interconnectedness of these factors, the company can identify more effective, holistic solutions that address the root causes of turnover rather than just the symptoms.

By following these steps and principles, systems thinking helps in developing a deeper understanding of complex problems and designing more effective and sustainable solutions.

Advantages of system thinking

System thinking is better than other approaches in understanding dependencies, interdependencies, and relationships in complex environments because it views the whole system as a network of elements that interact with each other, rather than examining individual components in isolation. This approach allows for a more accurate and comprehensive understanding of how changes in one part of the system affect the rest of the system. It helps in identifying patterns, trends, and behaviors over time, which can provide insights into the underlying structures that drive these observable patterns. By doing so, system thinking can highlight causal relationships that might otherwise remain unnoticed. This is because it focuses on the connections and interactions between components, which are often the source of emergent properties and complex behavior in systems. Therefore, system thinking is a powerful tool for understanding, managing, and changing complex systems.

Key rules of systems theory

The first law of systems theory - all dynamic systems are usually on the knife edge of falling apart.

The second law of systems theory - dynamic systems can stay that way for far longer than anyone might believe.

The third law of systems theory - when dynamic systems do finally break down they do catastrophically, usually at the worst possible time.

Key principles of system theory as per GPT-4o

  1. Interconnectedness: Everything is connected to everything else. The parts of a system do not function in isolation but affect each other in various ways.
  2. Holistic View: The system should be seen as a whole, not just as a collection of parts. Understanding the system requires looking at the interactions and relationships among the parts.
  3. Feedback Loops: Systems are regulated by feedback loops, which can be reinforcing (positive feedback) or balancing (negative feedback). These loops are crucial for system stability and change.
  4. Emergence: The properties and behaviors of a system are emergent, meaning they arise from the interactions and relationships among the parts, rather than from the parts themselves.
  5. Boundaries: Every system has boundaries that define what is inside and outside the system. Identifying and understanding these boundaries is essential for studying the system.
  6. Hierarchy: Systems are often nested within other systems. Each system can be a subsystem of a larger system and may contain subsystems within it.
  7. Dynamics: Systems are dynamic and change over time. Understanding a system involves studying how it evolves and adapts.
  8. Equilibrium and Disequilibrium: Systems can be in a state of equilibrium (stable) or disequilibrium (unstable). Systems tend to seek balance, but can also be disrupted by internal or external forces.
  9. Nonlinearity: Relationships within systems are often nonlinear, meaning that cause and effect are not directly proportional. Small changes can lead to large effects and vice versa.
  10. Self-Organization: Systems have the ability to self-organize and adapt without external control. This principle highlights the system's capacity for resilience and evolution.
  11. Delays: Time delays in feedback loops can lead to oscillations and affect the stability of the system. Recognizing delays is important for understanding system behavior.
  12. Purposefulness: Systems often have purposes or goals. Understanding the purpose of a system helps in understanding its behavior and outcomes.
  13. Diversity and Redundancy: Diversity within a system can enhance its resilience and adaptability. Redundancy (having multiple ways to achieve a function) can provide backup options and increase system stability.
  14. Leverage Points: Identifying leverage points—places within a system where a small change can lead to significant impacts—can be crucial for effectively managing and influencing systems.
System thinking
Pasted image 20240704101706.png
Pasted image 20240704101341.png

Knowledge graphs

Pasted image 20240702092012.png
  • human- and machine-readable
  • domain-specific
  • governed
  • interoperable

source of knowledge

Applying scientific thinking to AI projects

Scientific thinking

Because, when discussing a new project, a new problem you do this:

Medinorma.004.png
Digital twining.001.png

Because

  • we need to understand that dots that appear disconnected are connected
  • the sum of local solutions does not result in a solution to a global problem
Pasted image 20240222110236.png
Pasted image 20240222110225.png
Knowledge graphs.003.png
Knowledge graphs.002.png
Knowledge graphs.001.png

Science

Well-being

Pasted image 20240708172927.png

Guardrails to get fooled by

  • yourself
  • others

After 5000 years of human history, this is the best approach human kind has come up with to understand the world.

One can do everything with LLMs (Turing tarpit)
This is not a reason to use them to do everything

Pasted image 20240708172640.png
Buoy_light_in_storm.jpeg

Minimise pain

How to find answers to phenomenons we are not able to explain rationally (link logically a cause to a consequence)?

True or false do not matter;
what matter is that they make sense

How to cope with this digital deluge
and the
confusion, uncertainties, unwell-being
it contributes to create?

code-matrix.gif
Pasted image WallE.png

Do we want AI to

  • upgrade humans?
  • downgrade humans?

What do we want as

  • a society?
  • as jobs?
  • as food?
  • as AI?
  • as a future?

What do we (humans) want?

Making sense depends on one cognitive abilities and environment (e.g. culture, habits).
It varies from one individual to another

Pasted image Wasteland.png

Inability to explain mysterious phenomenons

such as

  • what is a disease?
  • why suffering?
  • what is death?
  • what happens after the body dies?

Cortisol: increase metabolic and triggers the immune response

Adrenaline: increases heart rate, expands air passages of the lungs, enlarges the pupils, and redistributes blood to the muscles

Noradrenaline: increase heart rate and blood pressure

From GPT-4

  1. Cortisol: This is the primary stress hormone. It increases sugars (glucose) in the bloodstream, enhances the brain's use of glucose, and increases the availability of substances that repair tissues. It also curbs functions that would be nonessential or detrimental in a fight or flight situation, such as immune system response, digestion, reproduction, and growth processes.

  2. Adrenaline (Epinephrine): This hormone is responsible for the immediate reactions we feel when stressed. It's the reason your heart beats faster, you breathe more rapidly, and your muscles tense. It also helps to convert glycogen into glucose in the liver, providing the body with extra energy.

  3. Norepinephrine: This hormone is often released into the bloodstream along with adrenaline. It increases arousal and alertness, speeds reaction time, and increases the capacity for action. It also enhances the formation and retrieval of memory and focuses attention.

  4. Aldosterone: This hormone helps regulate sodium and potassium levels in the body, which in turn regulates blood volume and blood pressure. It is released when the body is under stress to help maintain blood pressure.

  5. Glucagon: This hormone works with adrenaline to stimulate the liver to release glucose into the bloodstream, providing the body with the energy it needs to respond to stress.

  6. Growth Hormone: This hormone, released by the pituitary gland, also helps the body respond to stress. It stimulates protein synthesis and increases blood levels of glucose and fatty acids.

Stress hormones

IMG_5798.jpeg

From GPT-4

  1. Dopamine: Often referred to as the "feel-good" hormone, dopamine is associated with feelings of pleasure, satisfaction, and reward. It is released when we engage in activities that we find enjoyable, such as eating delicious food or having sex. Dopamine also plays a crucial role in motivating us to repeat these pleasurable activities.

  2. Oxytocin: Known as the "love hormone", oxytocin is released during physical touch, hugging, and sex. It promotes bonding, trust, and empathy, and is particularly important in romantic relationships and parent-child bonding. Oxytocin also helps reduce stress and anxiety.

  3. Endorphins: These are the body's natural painkillers. They are released during physical activity, laughter, and sex, helping to reduce pain and induce feelings of pleasure or euphoria. Endorphins are also responsible for the "runner's high" that people often experience after intense exercise.

  4. Serotonin: This hormone is associated with mood regulation, and higher levels of serotonin are linked to feelings of well-being and happiness. Serotonin also helps regulate sleep, appetite, and cognitive functions like memory and learning.

  5. Endocannabinoids: These are naturally produced by the body and bind to the same receptors as THC, the active ingredient in cannabis. They are involved in a variety of physiological processes including appetite, pain-sensation, mood, and memory. They also play a role in the feelings of relaxation and well-being.

  6. Phenylethylamine: This hormone is released in the brain when we fall in love and during sexual activity. It stimulates the body to release dopamine and norepinephrine, another hormone that can create feelings of joy.

  7. Ghrelin: Known as the "hunger hormone", ghrelin also stimulates the release of dopamine in the brain, and can therefore create feelings of pleasure and satisfaction when we eat.

Pleasure hormones

Too much,
too fast

Wasted people

Wastland

Pasted image 20240708171153.png
Pasted image 20240624142937.png
Pasted image 20240624142814.png
Pasted image 20240624142832.png
Pasted image Digitaly_wasted_people.png

Less essential and enjoyable activities

  • comfort, leisure
  • service society

Progress nowadays

is assimilated in a reductionist was to technology only

Industrial revolution.png

Technological progress

Pasted image 20240625122608.png
Pasted image 20240624142655.png
  • With consciousness came consciousness of self
  • With self-consciousness came consciousness of time and aging
  • With aging came the consciousness of self-end

Essential but painful and cumbersome physical activities:

  • hunting, gathering
  • physical work

Maximise pleasure

Pain & Pleasure

“Nature has placed mankind under the governance of two sovereign masters, pain and pleasure.” — Jeremy Bentham

Neurones connecting.mp4
‎DIKW variations.‎1001.png

with consciousness and reasoning came more pain: sorrows

Minimise pain

Pasted image 20240625122704.png

future gains

2023

Profit

Hope

Trust

  • markets, future (profits), progress
  • tomorrow will be better than today (thanks to technology)
‎Humans&AI.‎004.png

2021

Purpose of neurons/life is to connect to exchange information and create knowledge

Pasted image 20240708171820.png

What is rare has more value

Pasted image 20240703115123.png
Pasted image 20240703143519.png
Ancient_coins.png

Banks

Economy

Finance

First_US_bank.png
Pasted image 20240708171733.png

Consumerism

Productivism

Abundant and cheap energy

2012

Progress

A set of advancements in technology, science, and social organization efficiency

Pasted image 20240708171610.png

Materialism/Value/Capitalism

Question-answering chatbot

  • Draft answer for newly incoming questions from Health Authorities
  • Mimic a senior technical expert
  • Discuss with my data (document corpus or sets of tabular data)
  • ...

Automatic text writing

  • Automatically draft GxP deviations reports
  • Automatically draft report for Chemical Substance inventorying
  • Automatically draft Safety Data Sheets
  • ...
GenAI use cases for Pharma industry

Here comes Artificial Intelligence (AI)...

The cost of knowledge

Examples of AI use cases for the Pharma industry

Trade

Money

Social evolution, economy and finance

The notion of progress and its evolution

The business of Digital Transformation

Resulting digital complexity

Does it add value?

The scientific methodology is a systematic and logical approach used by scientists to explore phenomena, acquire new knowledge, or correct and integrate previous knowledge. Here are the key components and steps involved in this methodology:

  1. Make an observation and describe it: the initial stage where a scientist notices something of interest in the natural world. This can be anything from a specific event, pattern, behaviour, or anomaly. Observations are critical because they form the basis for asking questions and defining problems.
  2. Ask a question: begin by identifying a specific question or problem based your observations. This question should be clear and focused.
  3. Do background research: Gather existing information related to your question. This can involve reading scientific literature, reviewing previous studies, and consulting experts.
  4. Propose a hypothesis: A hypothesis is a tentative explanation or prediction that answers your question. It can be tested through experimentation. Formulate a hypothesis that provides a tentative answer A good hypothesis is specific, testable, and based on existing knowledge that was collected during the "background research" phase.
  5. Design and conduct experiments: Experimentation involves testing hypotheses through controlled and repeatable experiments. Plan and carry out experiments to test your hypothesis. Ensure that your experiments are controlled, repeatable, and account for variables. This step is crucial for investigating the relationships between variables and determining causality. Experiments are designed to be objective and to minimise bias, allowing for accurate and reliable results. Document all experiments using a standard documentation (metadata) methodology.
  6. Collect and analyze data: Gather data from your experiments and analyse it to determine whether it supports or refutes your hypothesis defined at step # 4. Use statistical methods where appropriate to interpret the data.
  7. Draw conclusions from experiments: Based on your data analysis, draw conclusions about your hypothesis. Determine whether it is validated, invalidated, or if further testing is required.
  8. Hypothesis validation or invalidation: Validation occurs when experimental results support the hypothesis, while invalidation occurs when results contradict it. Hypotheses must be falsifiable, meaning there should be a possible outcome that could prove them wrong.
  9. Communicate results: Share your findings with the scientific community through research papers, presentations, or other means. Peer review and publication are essential for validating and disseminating scientific knowledge.
  10. Refine and repeat: Science, more specifically Research is iterative. Based on your conclusions, you may need to refine your hypothesis, design new experiments, or explore additional questions. The process continues as new observations and data lead to further inquiry.

The scientific methodology is based on Empirical thinking

Advantages of scientific methodology

The scientific methodology is superior to other approaches in understanding the world because it is systematic, objective, and reproducible. It begins with a question or a hypothesis, followed by careful observation or experimentation. The data gathered is then analyzed and used to either support or reject the hypothesis. This process is unbiased, as it relies on empirical evidence and logical reasoning, not on personal beliefs or opinions. Furthermore, the results are verifiable, meaning other scientists can replicate the experiments to confirm the findings. This methodological rigor helps ensure the accuracy and reliability of scientific knowledge, making it a powerful tool for understanding the complex and diverse phenomena in the world.

Scientific methodology

1. Ask a questions:

  • What is the problem we are trying to solve?
  • What does success looks like?
  • How to measure it?

2. Do background research: Was it already solved and how?

In science and engineering, Root cause analysis (RCA) is a method of problem solving used for identifying the root causes of faults or problems. It is widely used in IT operations, manufacturing, telecommunications, industrial process control, accident analysis (e.g., in aviation, rail transport, or nuclear plants), medicine (for medical diagnosis), healthcare industry (e.g., for epidemiology), etc. Root cause analysis is a form of inductive (first create a theory [root] based on empirical evidence [causes]) and deductive (test the theory [underlying causal mechanisms] with empirical data) inference.

RCA can be decomposed into four steps:

  • Identify and describe the problem clearly
  • Establish a timeline from the normal situation until the problem occurs
  • Distinguish between the root cause and other causal factors (e.g., using event correlation)
  • Establish a causal graph between the root cause and the problem

RCA generally serves as input to a remediation process whereby corrective actions are taken to prevent the problem from recurring. The name of this process varies from one application domain to another. According to ISO/IEC 31010, RCA may include the techniques Five whys, Failure mode and effects analysis (FMEA), Fault tree analysis, Ishikawa diagram, and Pareto analysis.

Wikipedia

If you want to check something you want a reference

Root cause analysis

3. Propose a hypothesis:

On so on...

Can we reproduce it?

Let do it

Then

  • A LLM can do the job but...

Then

  • No need for a LLM

If

  • The solutions consists of inserting new values (numbers or text) in text templates

Or

  • There is no legacy artefacts that can serve as models

If

  • The solutions consists of the generation of text as a new narrative
  • There is a collection of legacy artefacts that can serve as text models

Why, why, why, why, why

based on initial success definition

MBC pitch for canvas.050.png

4. Design and conduct experiments: head to head comparison of models

Medinorma.021.png
‎Humans&AI.‎025.png
‎Humans&AI.‎026.png

Conclusion:

  • AI is a piece of technology; it is not inherently good/bad, stupid/intelligent (despite its name)
  • AI being truly "I" depends on people and usages: why, how and for what goal

  • Don't jump on using AI and ask yourself "Are my business/activities bound to quality, veracity, validation, regulation, (customer) trust?"
    - “No” then no problem, go ahead, use LLMs at will
    - “Yes”: wait a second, people are “fooled by their fluency but LLMs don’t understand how the world works” https://time.com/collection/time100-ai/6309052/yann-lecun/]. What they outputs are probabilities of word/pixel association based on biased training data

  • The world, especially in digital and AI times is not simple to understand. You need to harness this digital complexity if you want to get a competitive advantage
  • Be aware and take in account
    • how the brain work and its biases
    • how the neoliberal capitalism give a "profit" spin to everything it touches, especially technology
  • To keep you safe from falling in many pitfalls

Thank you for your attention !

  1. The opinions expressed in this presentation and on the following canvas are solely those of the presenter.

  2. Sources can point to commercial entities. I'm neither promoting these entities not their product or services. Just using digital material they produced to support the points of this course.

  3. Neither an historian nor an economist, nor a psychologist but just a biologist by training and now a DIK management expert with some generalist knowledge
Disclaimer
Pasted image Network_of_neurons.png

What does the brain "contains"

Brains are hard wired to understand stories

How does it acquire what it "contains"

Pasted image 20240624140321.png

Stories

A reassuring package

offering

  • answers to questions (e.g. solar deities )
  • solutions to problems (e.g. pray to be cured)
  • promises (e.g. better after-life, a messiah to save us)

How to explain phenomenons we don't understand?

Information + emotion = remembering

Medinorma.005.png

How our brain works

‎Humans&AI.‎001.png

Knowledge = now how to use information to solve/answer a problem/question

Content:

Part I (8 min) - Humans, their brains and ways of thinking

  • What make us what we are is our brain
  • What drives us is minimising pain and maximising pleasure
  • Gods, religions and messiahs

Part II (8 min) - History of progress: from Neolithic to Artificial Intelligence

  • The notion of progress and its evolution
  • Social evolution, economy and finance
  • The business of Digital Transformation
  • Resulting digital complexity
  • The cost of knowledge
  • Here comes Artificial Intelligence (AI)...
  • What do we (humans) want?

Part III (8 min) - How to cope with this digital deluge and the confusion, uncertainties, unwell-being it contributes to create?

  • Scientific thinking
  • System thinking

Conclusion

Abstract:
The belief in the advent of a messiah who acts as the saviour of a group of people is common to past and present societies. Whether it is identified as a religion, a cultural trend or a concept (e.g. the notion of progress), it always relies on the bio-chemical mechanisms underpinning brain functions, namely the way humans hope for a better future. In our digital times, there is a big hope that computational algorithms will solve/answer many problems/questions of moderns societies. While this is an obvious message from many big digital organisations and leaders, one can show healthy skepticism and ask simple questions: Are the problems AI is supposed to solve essential to human (well-)beings? What percentage of world population is benefiting from AI innovations? Is AI truly upgrading humans abilities? What is the part of pure commercial arguments in speeches promoting AI? Answering these questions is not easy. To best manage this complex topic, we propose to adopt methodologies that guarantee an iterative and guard-railed thinking process supporting a relatively safe and quick implementation of AI solutions. We will focus on generative AI applications and provide concrete examples supporting drug development and manufacturing processes.

a better future

Science

Hope

Pasted image 20240625122531.png
Pasted image 20240625122504.png

This is what drove brain evolution and the apparition of consciousness

Gods, religions and messiahs

What drives us is minimising pain and maximising pleasure

What makes us what we are is our brain

Pasted image 20240708173640.png
Pasted image 20240708173722.png
Pasted image 20240708173806.png
Pasted image 20240708173917.png
based on
but
reasoning capabilities of the neo-cortex
need to
but
Easy & quick: we invent
Long & difficult: we do
and bundle them as
and
helps to cope with
defined between 17th and 18th century
but started with
invented by
to ease tasks and remove
continued with
thanks to
enabling
enables
enables
by removing
and enabling to focus on
but
triggers
for
for
based on
in
leading to data mess
for
why?
for
A new
but the hype did not last long
and misuses are kicking in
we need some garde-rails
why?
why?
provide
applied to AI
but
but
GPUs
Main question to answer
using a predefined process
yes
no
and not this
why?
moreover
are a
integrated into business analysis
resulting in
combined with AI
business road
Is it compatible? Can we balance it?
human road