Scientific Reports,
Journal Year:
2023,
Volume and Issue:
13(1)
Published: April 8, 2023
Abstract
Misinformation
can
have
a
profound
detrimental
impact
on
populations’
wellbeing.
In
this
large
UK-based
online
experiment
(n
=
2430),
we
assessed
the
performance
of
false
tag
and
inoculation
interventions
in
protecting
against
different
forms
misinformation
(‘variants’).
While
previous
experiments
used
perception-
or
intention-based
outcome
measures,
presented
participants
with
real-life
posts
social
media
platform
simulation
measured
their
engagement,
more
ecologically
valid
approach.
Our
pre-registered
mixed-effects
models
indicated
that
both
reduced
engagement
misinformation,
but
was
most
effective.
However,
random
differences
analysis
revealed
protection
conferred
by
differed
across
posts.
Moderation
immunity
provided
is
robust
to
variation
individuals’
cognitive
reflection.
This
study
provides
novel
evidence
general
effectiveness
over
tags,
platforms’
current
Given
inoculation’s
effect
heterogeneity,
concert
will
likely
be
required
for
future
safeguarding
efforts.
Electronics,
Journal Year:
2021,
Volume and Issue:
10(11), P. 1348 - 1348
Published: June 5, 2021
In
recent
years,
we
have
witnessed
a
rise
in
fake
news,
i.e.,
provably
false
pieces
of
information
created
with
the
intention
deception.
The
dissemination
this
type
news
poses
serious
threat
to
cohesion
and
social
well-being,
since
it
fosters
political
polarization
distrust
people
respect
their
leaders.
huge
amount
that
is
disseminated
through
media
makes
manual
verification
unfeasible,
which
has
promoted
design
implementation
automatic
systems
for
detection.
creators
use
various
stylistic
tricks
promote
success
creations,
one
them
being
excite
sentiments
recipients.
This
led
sentiment
analysis,
part
text
analytics
charge
determining
polarity
strength
expressed
text,
be
used
detection
approaches,
either
as
basis
system
or
complementary
element.
article,
study
different
uses
analysis
discussion
most
relevant
elements
shortcomings,
requirements
should
met
near
future,
such
multilingualism,
explainability,
mitigation
biases,
treatment
multimedia
elements.
Nature Communications,
Journal Year:
2022,
Volume and Issue:
13(1)
Published: April 28, 2022
Interventions
that
shift
users
attention
toward
the
concept
of
accuracy
represent
a
promising
approach
for
reducing
misinformation
sharing
online.
We
assess
replicability
and
generalizability
this
prompt
effect
by
meta-analyzing
20
experiments
(with
total
N
=
26,863)
completed
our
group
between
2017
2020.
This
internal
meta-analysis
includes
all
relevant
studies
regardless
outcome
uses
identical
analyses
across
studies.
Overall,
prompts
increased
quality
news
people
share
(sharing
discernment)
relative
to
control,
primarily
intentions
false
headlines
10%
control
in
these
The
magnitude
did
not
significantly
differ
content
(politics
compared
with
COVID-19
related
news)
decay
over
successive
trials.
was
robustly
moderated
gender,
race,
political
ideology,
education,
or
value
explicitly
placed
on
accuracy,
but
larger
older,
more
reflective,
attentive
participants.
demonstrates
discernment.
Proceedings of the National Academy of Sciences,
Journal Year:
2021,
Volume and Issue:
119(1)
Published: Dec. 28, 2021
The
recent
emergence
of
machine-manipulated
media
raises
an
important
societal
question:
How
can
we
know
whether
a
video
that
watch
is
real
or
fake?
In
two
online
studies
with
15,016
participants,
present
authentic
videos
and
deepfakes
ask
participants
to
identify
which
which.
We
compare
the
performance
ordinary
human
observers
leading
computer
vision
deepfake
detection
model
find
them
similarly
accurate,
while
making
different
kinds
mistakes.
Together,
access
model's
prediction
are
more
accurate
than
either
alone,
but
inaccurate
predictions
often
decrease
participants'
accuracy.
To
probe
relative
strengths
weaknesses
humans
machines
as
detectors
deepfakes,
examine
machine
across
video-level
features,
evaluate
impact
preregistered
randomized
interventions
on
detection.
manipulations
designed
disrupt
visual
processing
faces
hinder
mostly
not
affecting
performance,
suggesting
role
for
specialized
cognitive
capacities
in
explaining
performance.
Science Advances,
Journal Year:
2021,
Volume and Issue:
7(36)
Published: Sept. 3, 2021
Professional
fact-checking,
a
prominent
approach
to
combating
misinformation,
does
not
scale
easily.
Furthermore,
some
distrust
fact-checkers
because
of
alleged
liberal
bias.
We
explore
solution
these
problems:
using
politically
balanced
groups
laypeople
identify
misinformation
at
scale.
Examining
207
news
articles
flagged
for
fact-checking
by
Facebook
algorithms,
we
compare
accuracy
ratings
three
professional
who
researched
each
article
those
1128
Americans
from
Amazon
Mechanical
Turk
rated
article’s
headline
and
lede.
The
average
small,
crowds
(i)
correlate
with
the
fact-checker
as
well
fact-checkers’
other
(ii)
predict
whether
majority
“true”
high
accuracy.
cognitive
reflection,
political
knowledge,
Democratic
Party
preference
are
positively
related
agreement
fact-checkers,
identifying
headline’s
publisher
leads
small
increase
in
fact-checkers.
