Between Predictive and Generative AI: Analogizing with Naïve Physics and Psychology
Overview
Naïve or folk physics and psychology are commonplace
in everyday human experience. For the former, it relates to common sense
reasoning or theories that help humans understand objects and events in our
world. For the latter, it helps in humans' understanding of different mental
processes, emotions, and behaviour about oneself and others.
This writing will attempt to consider analogically
naïve or folk physics and psychology as pedestals of the technological ingenuities
of predictive and generative Artificial Intelligence (AI) by following this
outline:
The above can be considered from the predictive prowess of David
Beckham who scored his famous goal from the halfway line, for naïve physics,
and the FBI profilers' ability to capture the Unabomber, for naïve psychology.
Those analogies will help us understand predictive AI and
generative AI, compare them to everyday experience, and bring to the fore their
benefits. Also, by analyzing these machine learning models, attempts will be
made to consider the role of humans in the artificial intelligence sphere. At
this point, Carl Jung’s collective unconscious is adopted and adapted to understand
why Africa as a continent is poor even when it has too many mineral/natural and
human resources to feed the entire world.
Halfway-line Goal!
This is Beckham’s signature goal that made him a
legend. That is the technique of curving the ball during flight once the goalkeeper
has stepped a little away from the goalposts and mentally calculating,
measuring both the impetus and velocity of his kick to ensure that the ball
gets into the goalposts before the goalkeeper can get back to his position.
His ability to predict the behaviour of the goalkeeper
was dependent on many factors, such as severally observing many goalkeepers who
stepped out of their penalty yard boxes and intuitively measuring the swiftness
of their ability to return to their goalposts. He achieved this because of an
understanding of the rules of nature. Beckham’s halfway-line goal seems like a
paradox that defies Zeno’s Paradox, and the concept of infinite divisibility.
Zeno’s Paradox
In the Tortoise and Achilles paradox, Zeno of Elea challenges
human intuition about time, motion, and infinity. In his paradox, Achilles, the
fastest hero created by Homer in the IIiad challenged the tortoise,
a slow-moving creature to a race. In this case, the tortoise argues that so
long as Achilles allows it a small head start, it will win the race. Achilles
trusting in his speed and strength agreed but he was not able to catch up or
overtake the tortoise because each time Achilles reached the tortoise's
starting point, the tortoise moved a little farther from him.
In this case, it can be said that Achilles’ prediction
that he would overtake the tortoise in an infinite divisibility race was flawed
because he never overtook the tortoise. However, Beckham’s prediction of target
divisibility was accurate, and he scored a goal. His ability to intuitively measure
and calculate on the flight, the distance between the position of the
goalkeeper when he was away from his goalposts, the impetus and velocity of an
inflated piece of rubber (football), and the force of his kick on the piece of
rubber to make it travel faster and quicker than the goalkeeper who was away
from his goalposts made him a legend of target divisibility.
FBI Profiler and the Unabomber
America’s Federal Bureau of Investigation (FBI)
profilers are notable for their ability to study, learn, and train their minds
to understand the behaviours of a criminal to capture them or predict their
next action. Often, they are credited with thinking like a criminal or articulating
through analyses the next move or subsequent developments of a crime.
The case immediately comes to mind is the Unabomber or
Ted Kaczynski as he was known. The Unabomber was a former professor of
mathematics who felt desperate and upset by modern inventions, such as
technology, and its impact on society. As such, he carried out a 16-year
bombing campaign of selected technology companies, universities, and airlines. That
drew the attention of the FBI who are charged with protecting Americans at home.
The FBI profilers put on their profiling hats and
started to collect data, information, and different writings of the Unabomber. They
studied, analyzed, and interpreted his writings and behaviour. Such analyses and
studies enabled them to identify patterns of behaviour, clues about movements,
and psychological traits. The analysis of the language structure of his
manifesto helped connect it to the Unabomber’s earlier writings. Therefore, his
writings and pattern of behaviour led to his arrest.
Let’s use the African continent as our next example
and in this case, we will try to find reasons or factors why Africa is a poor
continent. We will also ask at least three generative AI tools to respond to
the question of why Africa is poor.
Africa’s Political and Economic Archetypes
Since the late 50s and early 60s when African nations
got independence from their former colonial masters, nothing has changed in
terms of their political and economic positions in the world. As the old
colonial masters left, Africa inadvertently opened their doors to other masters:
lending institutions and lending nations. One begins to wonder. How could a
continent that ruthlessly fought for independence could not stand independent? One
cannot but assume that what is happening to Africa is not directed by either
the old colonial masters or the new masters found in lending institutions and
nations. Rather, Africa’s malady is controlled by the inner workings of their psyches.
If we ask Carl Jung, he will describe Africa’s problems as being affected by
primal drives that are wrapped in their unconscious.
Since its political independence, Africa has constantly
been interchanging between two types of evils: Soft and Hard. Soft evil is the endless
borrowing of money from international lending institutions and nations that has
led to the desperate dispersion of its people to every corner of the world. The
money borrowed from those bodies for the development of Africa has ended up in
the private pockets of their corrupt leaders and projects not worthy of their
names.
Hard evil is the privation of co-existence and
tolerance that now and again subject the continent to wars and violent
conflicts that decimate the population. It seems every African country identifies
with the myths and experiences of either soft or hard evil. Africa seems to
share an unexplainable commonality in this soft or hard evil. Jung would
describe these unconscious evils ripping Africa apart as their “collective
unconscious.” For Jung, “collective unconscious refers to the part of the
unconscious mind which is shared by all humans and is inherited from our
ancestors.”
If this collective unconscious is an inheritance from our
ancestors, politically African leaders are bewitched by the archetypes who have
remained their heroes’ past. African leaders stay in office until they are
dead. African leaders make themselves life-presidents. African leaders steal
from their people. They steal from themselves. African leaders use governance
as a means of enriching themselves, their families, and friends. These are
common features universally found among African leaders.
Bad governance, grift, and choice between soft and
hard evils are hereditary and are passed on from one generation to another. Their
political and economic archetypes can be construed as inherited malicious
behavioural patterns that are unconsciously applied daily.
When a new African leader takes office, gradually you
start to see the application of the unconscious use of archetypes. This ‘collective
unconscious’ is in every African psyche, and in every African psyche, there
lives the archetypes.
It is pertinent to argue at this point that Africa has
continued to generate the same types of leaders who have favoured the types of
leadership that surrender to either soft evil or hard evil. Africa is not
guided by forces outside but rather by the application of learned patterns following
early drives wrapped in the unconscious. To free Africa from these maladies,
there is the need for an introduction of a new learning model that redeems the
collective unconscious and compels Africa to face its ‘shadows,’ to use Jung’s
other imagery.
Let’s delve into predictive and generative AI and examine
how beneficial they are to humans and how human naïve physics and psychology
work together with the trained models of predictive and generative AI.
Between Predictive and Generative AI
Humans are in the middle between predictive and
generative AI because they are the scientists who develop the machines that eventually
become the predictors and generators of artificial intelligence. Therefore, a
brief consideration of the meaning of predictive and generative AI will help elucidate
why humans’ hope and epistemological certitude require a balance between human
intelligence and artificial intelligence.
Predictive AI
This is a method of data analysis that uses machine
learning (ML) to calculate some future event or outcome. Through analysis of
both current pertinent and historical data, predictive AI identifies patterns
and trends and thus forecasts outcomes or indicates daily or future likely
occurrences.
The predictive AI model is trained through supervised
learning to learn the patterns and relationships in a dataset of human-created
content. The model is presented with a set of human-created content and
corresponding labels from which it learns to generate content very similar to
the human-created content that bears the same labels.
This brief description of predictive AI exhibits words
like ‘calculate,’ ‘predict,’ ‘forecast,’ ‘patterns,’ and ‘trends.’ Such words
are not different in meaning as used in the naïve physics and psychology
examples given above: Beckham’s Half-way Goal line and FBI Profiler and the
Unabomber. The only difference is that naïve physics and psychology refer to
the untrained human perception of common phenomenal activities in the physical
world and the untrained human understanding of mental processes, and emotions, of
oneself and others.
Predictive AI is a trained model that has been
positioned to do things such as:
- To make recommendations to you based on your content-search preferences, interests, location, buying history, and feedback.
- To assist companies in optimising their inventory and supply chain management based on demand, distance, acceptability, cost, availability, and demography.
- To predict a country’s current and future financial status based on income and expenditure, governing patterns, credit history, borrowing habits, debt management, and currency value.
- To detect irregularities and threats in cybersecurity systems through the analysis of user behaviour, network activity, and threat intelligence.
- To generate information about the weather based on meteorological observations, patterns of wind movements, hot/cold air, geography, and atmospheric conditions.
These examples of predictive AI help humans in their
daily lives to improve their living conditions as well as protect them from
either cybercriminals or averting treacherous weather conditions. What about
generative AI? Let’s examine its meaning and why it has become a household name
in this age.
Generative AI
This is an artificial intelligence that can generate
text, images, code, or any type of content according to the prompt of the user.
Just like predictive AI, a generative AI model is trained through supervised
learning to learn the patterns and relationships in a dataset of human-created
content. It does this by learning from existing data and can generate new data
very similar in style to the training data.
The difference between predictive and generative AI is
while the former outputs predictions and forecasts, the latter outputs new
content. Generative AI has quite some benefits:
- It can explore large unstructured data.
- It can help to improve customer interactions.
- It does well in repetitive tasks and can help organizations devote to other non-repetitive tasks.
- It has made research easier and convenient by bringing research materials to hand.
Generative AI is meant for humans and not humans for
generative AI. How do we position this essential and helpful tool so that it equitably
serves our interests? In the next section and adopting Jung’s basic principle
of ‘collective unconscious’ an attempt will be made to describe generative AI
from the perspective of ‘generative unconscious.’
Generative Unconscious
The idea of generative unconscious stems from Jung’s psychotherapeutic
analysis based on his research on the collective unconscious, where the
commonly found symbols and myths form part of the universally shared collective
unconscious. Above, the same collective unconscious was used to describe
Africa’s political and economic archetypes, where different layers of inherited
memories form the totality of their experience.
The generative unconscious is placing the generative AI
tools to the test. How would these tools generate responses to a single
question? Africa is the case in point because it has been used as an example of
a continent whose collective unconscious controls and motivates the inner
workings of their minds.
Three generative AI tools are chosen for this
exercise: Copilot, Google Bard, and ChatGPT. One simple and same question was
asked each of the tools: Why is Africa poor? The responses from each one
of them were nothing far from what might be described as a ‘generative unconscious.’
These generative AI tools listed some reasons why
Africa is poor. However, it advised that the details of the information may not
be correct and that it is the user’s responsibility to check for the veracity
of the responses. Here, it is not the veracity of the responses of the AI tools
that are being considered but their capacity to generate responses that suggest
a type of collective unconscious, as if the generative AI tools generate
content from the memories of different data.
Copilot
Copilot gives these reasons or factors that
contributed to Africa’s poverty:
Historical and Colonial Legacy
It then listed the following as some of the factors
that contributed to poverty in Africa.
- History of slavery
- Colonization
- Exploitation by foreign powers
- Weak institutions
- Poor governance
- Unequal distribution of resources
Economic Challenges and Unemployment
It listed some contributing factors in that regard:
- Low productivity
- High inflation
- Debt
- Corruption
- Trade barriers
- Unemployment
Health Issues and Access to Medical Care
It listed some contributing factors as:
- Lack of adequate healthcare systems and infrastructure
- Covid19 pandemic
- Vulnerability to preventable diseases
- Malnutrition
- Maternal and child mortality
Climate Change and Environmental Degradation
- Droughts
- Floods
- Desertification
- Crop failures
- Food security
- Water availability
- Deforestation
- Soil erosion
- Pollution
Bard
Broadly, Bard categorized Africa’s poverty into
historical, economic, political, and social influences.
Historical Factors
- The Legacy of Colonialism
- Forceful extraction of resources and wealth from Africa by colonial powers
- Imposition of their political and economic systems on Africa
- The Cold War
- The division of Africa and drawn into the proxy conflicts between the United States and the Soviet Union
Economic Factors
- Trap in a cycle of poverty
- Dependence on primary exports
- Lack of infrastructure – good roads, electricity, and sanitation
Political Factors
- Corruption
- Weak institutions – the police, judiciary, and the civil service
- Political instability
- Coups, civil wars, and ethnic conflicts.
Social Factors
- High population growth
- Disease – HIV/AIDS, and malaria
- Gender inequality
- Limitation of women's participation in the economy, access to education and healthcare.
ChatGPT
This tool stated that several key reasons contribute
to Africa’s economic challenges:
Historical Factors
- Colonialism
- Governance systems
- Social structures
- European powers' exploitation of Africa’s resources
- Their disruption of local economies
- Establishment of extractive systems that persisted even after Africa’s independence.
Economic Challenges
- High levels of debt
- Limited access to capital
- Inadequate infrastructure
- Reliance on primary commodities for export
- Fluctuating commodity prices
- Limited value addition
Governance and Corruption
- Weak institutions
- Siphoning of resources
Conflict and Instability
- Conflicts and civil wars
- Displacement of people
- Destruction of infrastructure
- Social unrest
Health Challenges
- Diseases – HIV/AIDS, malaria, other health issues
Environmental issues
- Droughts
- Desertification
- Climate change
- Food security
- Access to water
Limited Access to Education and Skills
- Inadequate access to quality education
- Limited access to skills training
Global Trade Dynamics
- Unfair trade practices
- Market barriers
- Unequal terms of trade with developed countries
The three generative AI tools used to get responses to
the question: Why is Africa poor gave three almost similar responses
notwithstanding the uniqueness of each generative AI tool. This suggests that
they derive their responses from the same pool of knowledge or episteme, though
trained using different models.
Summary
The key points raised in this writing are that naïve or
folk physics and psychology influence our everyday human experience.
Through mental calculation and measuring both the impetus
and velocity of his kick to ensure that the ball gets into the goalposts before
the goalkeeper can get back to his position, David Beckham established a
halfway-line goal signature.
Also, through accurate profiling, the FBI identified
patterns of behaviour, clues about movements, and psychological traits of the
Unabomber.
Using Carl Jung’s analyses of collective unconscious, there
is a suggestion that Africa is bewitched by the archetypes who have remained
their heroes’ past.
Finally, a question about: Why Africa is a poor
continent was asked three generative AIs: Copilot, Bard, and ChatGPT. The
responses by each were so similar that nothing far from what might be described
as ‘generative unconscious.’
Sources
Jung, C.G. (1968). The Archetypes and the Collective
Unconscious (R.F.C. Hull, Trans.; 2nd ed.). Routledge.
“Generative AI vs. Predictive AI” in SS&C Blueprism,
10 October 2023, Colin Redbond. https://www.blueprism.com/resources/blog/generative-ai-vs-predictive-ai/
accessed 30/11/2023.
“Unlocking The Power of Predictive Analytics With AI”,
in Forbes, 11 August 2021, Kevin Beasley. https://www.forbes.com/sites/forbestechcouncil/2021/08/11/unlocking-the-power-of-predictive-analytics-with-ai/?sh=380e7b0d6b2a
accessed 25/11/2023.
“Predictive AI in Cybersecurity: What Works and How to Understand It”, in Blackberry Blog, 18 October 2023, Shiladitya Sircar. https://blogs.blackberry.com/en/2023/10/predictive-ai-in-cybersecurity accessed 25/11/2023.
Comments