Postphenomenological approach to smart education: ChatGPT and technological mediation

Abstract

Students in the twenty-first century are different from the students of the past. Traditional education methodologies become less and less efficient while old learning models are rapidly replaced by new approaches from the Smart-education domain (such as personalized learning, game-based learning, etc). A significant part of contemporary learning approaches is played by novel AI adaptive educational systems. And now, LLMs (like ChatGPT) are changing the educational landscape. These systems aim to provide students with highly personalized educational suggestions, might adapt to specific learning needs, and can improve recommendations within the course through machine learning algorithms.
However, we don't know much about the non-neutrality of these systems. Here I propose a postphenomenolgical framework that might help to assess the non-neutrality of ChatGPT on three levels: perceptual, existential and moral.

Full text

Dmytro Mykhailov
Southeast university, Nanjing (China)
SPT2023
Tokyo
Postphenomenological approach to
smart education
AI and technological mediation

Content
•Part 1. Discussion and important
elements of ChatGPT
•Part 2. Benefits and Threads ChatGPT
can have for the smart education
•Part 3. Postphenomenological
perspective on non-neutrality of
ChatGPT
•Instead of conclusion. ChatGPT and
new opportunities for
postphenomenology

Discussions around ChatGPT
Main components of ChatGPT
New communicative situation
Part 1
Important elements of ChatGPT

Starting points
•Although the debates about canceling ChatGPT are still on, we need to learn
how to deal with ChatGPT because now this is a part of our collective
future.
•To know how to deal with this technology we need to understand its both
benefits and drawbacks that the system can create.
•From postphenomenological perspective we need in order to better
understand ChatGPT we have to start with human-technology relations
and so to see how the technology is non-neutral.

How the model is being designed?
•Neural networks + DATA - These machine language systems consist
of large artificial neural networks and are “trained” via a trial-and-
error process over mountains of data.
•The neural networks are rewarded when their algorithmically
generated words or images successfully meet the requirements of
the designer.
•After training, these generative models can create entirely new
synthetic artifacts (e.g. text in natural language; images, songs, etc.).

ChatGPT as a tutor
New communicative situation
•Understands the natural language
•Due to the sophisticated machine learning
algorithms ChatGPT creates a highly
personalized communication and
permanently improves itself.
•ChatGPT is permanently accessible
•The amount of the information that
ChatGPT is able to share is incomparable
to any other human being (a group)
•No bulling (under the normal
circumstances if the model operates
properly).

The request: Explain quantum physics like a fairy-tail for a small kid.
ChatGPT

Main components of
ChatGPT
•The student does not need to generate
text word by word anymore
•The new principles of the prompting
engineering such as: (1) using precise
prompts; (2) create many iteration cycles
to get the concrete answer to the question
(3) personalize the model according to
your own needs by using right prompts
•For example, ChatGPT has different
personas the user can switch between.
One can also ask ChatGPT to write like a
human.

Limitations of ChatGPT
•Sometimes ”hallucinates” things
•Provides inaccurate information and occasionally
fails simple logical reasoning tasks such as math
questions
•Can be biased and so generates a biased content
•Might cause environmental issues

What is smart education?
ChatGPT and benefits for educators
ChatGPT and benefits for learners
ChatGPT and possible threads
Part 2
Benefits and Threads ChatGPT can have
for the smart education

What is Smart Education?

ChatGPT and benefits for students
•Conversational AI. As a LLM optimized for dialogue, ChatGPT helps users test their
ideas, and reflect upon their thinking through receiving conversational feed- back.
•Language translator. Users can translate text back and forth from various languages,
or simplify text, which can be used to, for example, read otherwise incomprehensible
text.
•Code/Text generator. Similarly to text generation, ChatGPT is able to produce
programming code in multiple languages.
•Communication improvement. Students can use ChatGPT to fix the grammar in
their writing. ChatGPT can also more holistically improve the communication by
reorganizing sentences and entire paragraphs, and suggesting words and phrases.
Main capabilities of ChatGPT

ChatGPT and benefits for students
•Learn new languages.
•Virtual Simulation: ChatGPT can take on the persona of historical
figures, scientists, or literary characters, allowing students to engage
in virtual simulations and role-playing activities. This immersive
approach can deepen their understanding of various subjects and
enhance critical thinking
Main capabilities of ChatGPT

Perceptual non-neutrality
Existential non-neutrality
Moral non-neutrality
Part 3
Postphenomenological perspective on non-
neutrality of ChatGPT

Three levels of non-neutrality
Perceptual level
ChatGPT mediates the student’s
attention
Existential / Practical level
ChatGPT changes the learning
practices
The moral level
ChatGPT is changing who is a
good student

Perceptual level
The model mediates the
student’s attention
Increased Engagement: ChatGPT's
interactive nature can make learning more
engaging for students. Interacting with an AI-
powered chatbot can create a sense of
novelty and excitement, capturing
students' attention and fostering a
positive perception of the learning
experience.
Reduced Intimidation: Some students may
find it less intimidating to interact with
ChatGPT compared to asking questions or
seeking help from a teacher or peers directly.
This can create a more comfortable learning
environment

Existential / Practical level
The model changes the learning
practices
Accessible and On-Demand Support:
With ChatGPT, students have access to on-
demand assistance outside of classroom.
This can help address their questions and
concerns promptly, foster independent
learning, and create a sense of continuity in
the learning process.
Data-Driven Insights: ChatGPT's
interactions with students can generate
valuable data and insights that educators
can analyze. This data can provide feedback
on students' learning patterns,
misconceptions, and areas for
improvement, enabling educators to make
data-driven instructional decisions.

The moral level
The model is changing who is a
good student
Shifting Focus from Memorization to Problem-
Solving: ChatGPT's ability to provide on-demand
information may reduce the emphasis on
memorization as a measure of a good student. Instead,
problem-solving skills, critical thinking, and the ability
to apply knowledge become more important, as
students can access information easily and focus on
higher-order cognitive skills.
Individualized Learning Paths: ChatGPT's capacity
for personalized learning can challenge the notion of a
one-size-fits-all approach to education. A good
student may no longer be solely defined by their ability
to conform to a standardized curriculum. Instead,
success may be measured by a student's ability to
navigate their own unique learning path and achieve
mastery in areas that align with their interests and
strengths.

Instead of conclusion
ChatGPT and new opportunities for
postphenomenology

New possibilities
1. ChatGPT and Multistability
2. Text-to-image generative models and the future of the
image
3. Active passivity and ChatGPT
4. Technological mediation, ChatGPT and Transcendence

Full- text available as a part of
Springer Nature Content Sharing
Initiative via oicial Springer
SharedIt Link:
https://rdcu.be/c5JeR

Dmytro Mykhailov: 101300137@seu.edu.cn
Thank you for your attention