Disciplinary and Interdisciplinary Science Education Research,
Год журнала:
2025,
Номер
7(1)
Опубликована: Март 31, 2025
Abstract
Advancements
in
artificial
intelligence
(AI),
particularly
generative
AI
models
such
as
ChatGPT,
offer
transformative
opportunities
to
enhance
educational
practices
STEM
disciplines.
Thermodynamics,
a
fundamental
subject
engineering
education,
presents
significant
challenges
due
its
abstract
nature
and
common
misconceptions.
This
study
investigates
the
effectiveness
of
integrating
ChatGPT
supplemental
pedagogical
tool,
guided
by
constructivist
inquiry-based
approach
using
Constructivist
Inquiry-Based
Learning
Prompting
(CILP)
framework,
conceptual
understanding
address
misconceptions
an
introductory
thermodynamics
course
for
first-year
Moroccan
students.
A
quasi-experimental
design
was
used,
with
120
students
equally
divided
into
control
experimental
groups.
The
group
received
traditional
instruction,
whereas
ChatGPT-assisted
instruction.
Conceptual
measured
pre-
post-tests,
while
student
perceptions
acceptance
were
collected
via
weekly
surveys.
Results
showed
that
significantly
outperformed
group,
exhibiting
greater
improvements
reduction
qualitative
misconceptions,
related
entropy
internal
energy.
However,
some
quantitative
persisted,
underscoring
ChatGPT’s
limitations
advanced
reasoning
tasks,
problem-solving,
numerical
calculations.
Students
reported
high
satisfaction
usability
instructional
support.
Moreover,
targeted
use
rather
than
frequent
reliance,
correlated
optimal
learning
outcomes.
These
findings
underscore
potential
education
within
inquiry-based,
environments
provide
evidence
effective
integration
tools
improve
outcomes,
resource-constrained
settings.
Education Sciences,
Год журнала:
2025,
Номер
15(3), С. 280 - 280
Опубликована: Фев. 24, 2025
Large
language
model
(LLM)
tools,
such
as
ChatGPT,
are
rapidly
transforming
engineering
education
by
enhancing
tasks
like
information
retrieval,
coding,
and
writing
refinement,
which
critical
to
the
problem-solving
technical
focus
of
disciplines.
This
study
investigates
how
students
use
LLM
tools
challenges
they
face,
offering
insights
into
adoption
AI
technologies
in
academic
settings.
A
survey
539
from
12
leading
Chinese
universities,
using
UTAUT
framework,
examines
factors
technological
expectations,
environmental
support,
personal
characteristics.
The
key
findings
include
following:
(1)
Over
40%
with
18.8%
regarding
them
indispensable.
(2)
Trust
AI-generated
content
remains
a
central
challenge,
must
critically
evaluate
its
accuracy
reliability.
(3)
Environmental
support
significantly
affects
usage,
notable
regional
disparities,
particularly
between
eastern
other
regions
China.
(4)
persistent
digital
divide,
influenced
gender,
level,
socioeconomic
background,
depth
effectiveness
tool
use.
These
results
underscore
need
for
targeted
address
demographic
disparities
optimize
integration
education.
Disciplinary and Interdisciplinary Science Education Research,
Год журнала:
2025,
Номер
7(1)
Опубликована: Март 31, 2025
Abstract
Advancements
in
artificial
intelligence
(AI),
particularly
generative
AI
models
such
as
ChatGPT,
offer
transformative
opportunities
to
enhance
educational
practices
STEM
disciplines.
Thermodynamics,
a
fundamental
subject
engineering
education,
presents
significant
challenges
due
its
abstract
nature
and
common
misconceptions.
This
study
investigates
the
effectiveness
of
integrating
ChatGPT
supplemental
pedagogical
tool,
guided
by
constructivist
inquiry-based
approach
using
Constructivist
Inquiry-Based
Learning
Prompting
(CILP)
framework,
conceptual
understanding
address
misconceptions
an
introductory
thermodynamics
course
for
first-year
Moroccan
students.
A
quasi-experimental
design
was
used,
with
120
students
equally
divided
into
control
experimental
groups.
The
group
received
traditional
instruction,
whereas
ChatGPT-assisted
instruction.
Conceptual
measured
pre-
post-tests,
while
student
perceptions
acceptance
were
collected
via
weekly
surveys.
Results
showed
that
significantly
outperformed
group,
exhibiting
greater
improvements
reduction
qualitative
misconceptions,
related
entropy
internal
energy.
However,
some
quantitative
persisted,
underscoring
ChatGPT’s
limitations
advanced
reasoning
tasks,
problem-solving,
numerical
calculations.
Students
reported
high
satisfaction
usability
instructional
support.
Moreover,
targeted
use
rather
than
frequent
reliance,
correlated
optimal
learning
outcomes.
These
findings
underscore
potential
education
within
inquiry-based,
environments
provide
evidence
effective
integration
tools
improve
outcomes,
resource-constrained
settings.
