Electronics,
Journal Year:
2024,
Volume and Issue:
13(15), P. 2947 - 2947
Published: July 26, 2024
Deepfake
technology
uses
artificial
intelligence
to
create
realistic
but
false
audio,
images,
and
videos.
poses
a
significant
threat
the
authenticity
of
visual
content,
particularly
in
live-stream
scenarios
where
immediacy
detection
is
crucial.
Existing
approaches
have
limitations
challenges,
prompting
need
for
more
robust
accurate
solutions.
This
research
proposes
an
innovative
approach:
combining
eye
movement
analysis
with
hybrid
deep
learning
model
address
real-time
detection.
The
proposed
integrates
two
neural
network
architectures,
MesoNet4
ResNet101,
leverage
their
respective
architectures’
strengths
effective
classification.
lightweight
CNN
designed
explicitly
detect
subtle
manipulations
facial
images.
At
same
time,
ResNet101
handles
complex
data
feature
extraction.
Combining
localized
deeper,
comprehensive
representations
our
achieves
enhanced
performance
distinguishing
between
manipulated
authentic
videos,
which
cannot
be
conducted
naked
or
traditional
methods.
evaluated
on
diverse
datasets,
including
FaceForensics++,
CelebV1,
CelebV2,
demonstrating
compelling
accuracy
results,
attaining
0.9873
0.9689
0.9790
showcasing
its
robustness
potential
real-world
deployment
content
integrity
verification
video
forensics
applications.
Psychology and Marketing,
Journal Year:
2024,
Volume and Issue:
41(4), P. 880 - 898
Published: Jan. 4, 2024
Abstract
While
consumer
engagement
(CE)
in
the
context
of
artificially
intelligent
(AI‐based)
technologies
(e.g.,
chatbots,
smart
products,
voice
assistants,
or
autonomous
cars)
is
gaining
traction,
themes
characterizing
this
emerging,
interdisciplinary
corpus
work
remain
indeterminate,
exposing
an
important
literature‐based
gap.
Addressing
gap,
we
conduct
a
systematic
review
89
studies
using
Preferred
Reporting
Items
for
Systematic
reviews
and
Meta‐Analyses
(PRISMA)
approach
to
synthesize
AI‐based
CE
literature.
Our
yields
three
major
CE,
including
(i)
Increasingly
accurate
service
provision
through
;
(ii)
Capacity
(co)create
consumer‐perceived
value
,
(iii)
CE's
reduced
effort
their
task
execution
.
We
also
develop
conceptual
model
that
proposes
antecedents
personal,
technological,
interactional,
social,
situational
factors,
consequences
consumer‐based,
firm‐based,
human‐AI
collaboration
outcomes.
conclude
by
offering
pertinent
implications
theory
development
future
research
questions
derived
from
proposed
CE)
practice
reducing
costs
brand/firm
interactions).
Journal of Retailing and Consumer Services,
Journal Year:
2024,
Volume and Issue:
79, P. 103831 - 103831
Published: March 26, 2024
The
surge
in
social
media's
popularity
has
catalyzed
the
emergence
and
growth
of
media
influencers
(SMIs)
their
pivotal
role
marketing
strategy.
This
study
examines
impact
SMI
over-endorsement
on
consumer
purchase
intentions,
mediating
this
relationship
through
influencer
authenticity
credibility,
moderating
it
with
product
interest.
Employing
a
moderated-mediation
model
using
partial
least
squares
structural
equation
modeling
(PLS-SEM),
analyzes
survey
data
collected
from
303
users.
Contrary
to
intuitive
expectations,
does
not
directly
affect
intentions;
instead,
its
negative
influence
is
significantly
mediated
by
reduced
perceptions
credibility
among
SMIs,
interest
serving
as
mitigating
mechanism.
revelation
challenges
conventional
beliefs
revealing
that
repercussions
excessive
endorsements
are
erosion,
overshadowing
any
concerns
regarding
authenticity.
Importantly,
adverse
effect
can
be
mitigated
fostering
strategic
buffer.
suggests
rather
than
curtailing
endorsement
frequencies,
brands
SMIs
should
focus
amplifying
(e.g.,
enhancing
engagement
relevance)
thus
sustain
or
even
increase
activities
without
compromising
perceived
influencers,
thereby
maintaining
effectiveness
campaigns.
therefore
contributes
novel
insights
into
dynamics
marketing,
particularly
effects
behavior
conditions
under
which
these
may
neutralized.
Psychology and Marketing,
Journal Year:
2024,
Volume and Issue:
41(8), P. 1833 - 1853
Published: May 2, 2024
Abstract
Virtual
influencers
have
great
potential
to
shape
the
audience's
perception
and
motivate
behavioral
change.
Drawing
upon
anthropomorphism
theory
social
identity
theory,
this
paper
examined
interaction
effect
of
racial
homophily
virtual
in
shaping
audiences'
trust
subsequently
pro‐environmental
behaviors
low‐
high‐cost
contexts.
Two
online
studies
employing
experimental
materials
generated
by
artificial
intelligence
were
conducted.
Study
1
revealed
that
high
(human‐like
influencers)
(local‐like
are
more
effective
for
eliciting
context
low‐cost
behaviors.
The
findings
also
showed
low
(foreign‐like
would
further
weaken
level
a
influencer
(animate‐like
influencers).
2
aimed
replicate
from
enhanced
influencer.
moderated
mediation
analyses
confirmed
hypothesized
relationships
proposed
research
model.
This
advanced
our
understanding
on
moderating
role
serves
as
cognitive
shortcut
quick
decisions.
However,
(foreign‐like)
be
symbolic
These
insights
provide
valuable
guidance
businesses,
marketers,
environmental
advocates
aiming
leverage
encouraging
sustainability
practices.
