Philosophical Psychology,
Год журнала:
2024,
Номер
unknown, С. 1 - 27
Опубликована: Сен. 3, 2024
Conversational
agents
(CA)
are
thought
to
be
promising
for
psychotherapy
because
they
give
the
impression
of
being
able
engage
in
conversations
with
human
users.
However,
given
high
risk
therapy
patients
who
already
a
vulnerable
situation,
there
is
need
investigate
extent
which
CA
contribute
goals
and
discuss
CA's
limitations,
especially
complex
cases.
In
this
paper,
we
understand
as
way
dealing
existential
situations
position
CAs
context
therapeutic
experience
patients.
This
determined
by
patient's
unique
personal
specific
goals.
We
suggest
that
fundamentally
dialogical
activity,
it
crucially
involves
work
on
self
one's
self-narrative.
brings
us
our
central
question:
possible
productive
dialogue,
their
limitations
epistemic
agents?
will
several
those
show
how
these
undermine
possibility
engaging
illustrate
through
discussions
cases
grief
abuse.
Behaviour and Information Technology,
Год журнала:
2023,
Номер
unknown, С. 1 - 41
Опубликована: Ноя. 27, 2023
This
study
presents
a
systematic
literature
search
and
narrative
meta-review
of
the
current
state
research
on
conversational
agents
(CAs),
including
embodied
CAs,
chatbots,
social
assistive
robots
(SARs).
The
investigation
identifies
1,830
academic
articles,
which
315
articles
satisfied
inclusion
criteria
for
review.
Systematic
reviews
across
various
fields
are
reported,
mental
disorders,
neurodevelopmental
dementia/cognitive
impairment,
other
medical
conditions,
elderly
support,
health
promotion,
health,
education,
industrial
applications,
agent
characteristics,
robot
characteristics.
highlights
challenges
in
CA
research,
such
as
scarcity
high-quality
comparative
studies
acceptance
CAs
by
users
caregivers,
particularly
support.
article
also
categorises
ethical
discussions
into
nine
elements:
privacy,
safety,
innovation,
user
acceptance,
psychological
attachment,
care
philosophy,
evaluation,
systems
compatibility,
rule
development.
It
offers
insights
development
future
guidelines.
role
fostering
human
relationships
through
their
function
is
emphasised
to
provide
guidance
subsequent
implementation.
As
advancements
technology
continue
progress,
there
an
increasing
demand
sophisticated
investigations
addressing
relationships,
emotions,
self.
Science & Technology Libraries,
Год журнала:
2023,
Номер
43(3), С. 225 - 234
Опубликована: Сен. 8, 2023
ABSTRACTChatGPT
is
becoming
more
and
popular
in
the
healthcare
sector
as
a
way
to
deliver
quick
easy
access
health
information.
However,
little
known
about
how
participants
view
ChatGPT
when
seeking
To
find
participants'
perspectives
on
using
for
information
pinpoint
areas
where
ChatGPT's
functionality
design
should
be
improved.
Sixteen
individuals
who
had
used
look
underwent
semi-structured
interviews.
Themes
were
found
via
thematic
data
analysis.
was
use
handy
getting
They
did,
however,
voice
reservations
over
dependability
credibility
of
complicated
medical
situations.
Additionally
highlighted
essential
requirements
while
emotional
support
empathy.
For
professionals
Chatbot
developers,
study's
findings
have
both
theoretical
real-world
applications.
The
study
highlights
necessity
create
with
accessibility,
ease,
trustworthiness,
personalization,
usability
mind
underscores
significance
person-centered
approach
healthcare.
has
potential
improve
behavior.
there
constant
demand
advancements.KEYWORDS:
Artificial
intelligenceChatbotsChatGPThealth
behaviorqualitative
research
Disclosure
statementNo
conflict
interest
reported
by
author(s).
Journal of Medical Internet Research,
Год журнала:
2023,
Номер
25, С. e50767 - e50767
Опубликована: Сен. 29, 2023
Conversational
agents
(CAs),
or
chatbots,
are
computer
programs
that
simulate
conversations
with
humans.
The
use
of
CAs
in
health
care
settings
is
recent
and
rapidly
increasing,
which
often
translates
to
poor
reporting
the
CA
development
evaluation
processes
unreliable
research
findings.
We
developed
published
a
conceptual
framework,
designing,
developing,
evaluating,
implementing
smartphone-delivered,
rule-based
conversational
agent
(DISCOVER),
consisting
3
iterative
stages
design,
development,
implementation,
complemented
by
2
cross-cutting
themes
(user-centered
design
data
privacy
security).This
study
aims
perform
in-depth,
semistructured
interviews
multidisciplinary
experts
share
their
views
on
definition
classification
evaluate
validate
DISCOVER
framework.We
conducted
one-on-one
via
Zoom
(Zoom
Video
Communications)
12
using
an
interview
guide
based
our
framework.
were
audio
recorded,
transcribed
team,
analyzed
thematic
analysis.Following
participants'
input,
we
defined
as
digital
interfaces
natural
language
engage
synchronous
dialogue
≥1
communication
modality,
such
text,
voice,
images,
video.
classified
13
categories:
response
generation
method,
input
output
modalities,
purpose,
deployment
platform,
appearance,
length
interaction,
type
CA-user
initiation,
style,
personality,
human
support,
intervention.
Experts
considered
framework
could
be
adapted
for
artificial
intelligence-based
CAs.
However,
despite
advances
intelligence,
including
large
models,
technology
not
able
ensure
safety
reliability
settings.
Finally,
aligned
feedback,
present
updated
iteration
(CHAT)
key
considerations
themes:
ethics,
user
involvement,
security.We
expanded,
validated
CHAT
aim
at
guiding
researchers
from
variety
backgrounds
different
levels
expertise
implementation
Journal of Medical Internet Research,
Год журнала:
2024,
Номер
26, С. e53829 - e53829
Опубликована: Дек. 6, 2024
Background
Health
promotion
and
growth-based
interventions
can
effectively
improve
individual
well-being;
however,
significant
gaps
in
access
utilization
still
exist.
Objective
This
study
aims
to
develop
test
the
effectiveness
implementation
of
a
new,
widely
targeted
conversational
agent
prevention
program
(Zenny)
designed
enhance
well-being.
Methods
A
total
1345
individuals
United
States
were
recruited
online
randomly
assigned
either
(1)
self-help
intervention
delivered
via
an
automated
on
WhatsApp
or
(2)
active
control
group
that
had
evidence-based
wellness
resources
available
online.
The
primary
outcomes
well-being
(measured
using
5-item
World
Organization
Well-being
Scale),
psychosocial
flourishing
(assessed
with
Flourishing
positive
psychological
health
(evaluated
Mental
Continuum-Short
Form).
Outcome
measures
collected
at
baseline
again
1
month
postassessment.
All
analyses
conducted
intention-to-treat
approach.
Results
Both
groups
showed
improvements
(self-help
effect
size:
Cohen
d=0.26,
P<.001;
d=0.24,
P<.001),
(intervention:
d=0.19,
control:
d=0.18,
d=0.17,
P=.001;
P<.001)
However,
there
no
differences
between
2
(P
ranged
from
.56
.92).
As
hypothesized
priori,
greater
number
days
spent
actively
engaging
was
associated
larger
postassessment
among
participants
(β=.109,
P=.04).
Conclusions
findings
this
suggest
free
as
effective
web
resources.
Further
research
should
explore
strategies
increase
participant
engagement
over
time,
only
portion
involved,
higher
linked
Long-term
follow-up
studies
are
also
necessary
assess
whether
these
effects
remain
stable
time.
Trial
Registration
ClinicalTrials.gov
NCT06208566;
https://clinicaltrials.gov/ct2/show/NCT06208566;
OSF
Registries
osf.io/ahe2r;
https://doi.org/10.17605/osf.io/ahe2r
Behaviour and Information Technology,
Год журнала:
2024,
Номер
unknown, С. 1 - 15
Опубликована: Март 27, 2024
While
chatbots
show
promise
for
large-scale
mental
health
screening,
few
offer
interactive,
free-text
conversations,
limiting
their
appeal
self-administered
screening
and
impeding
the
timely
detection
of
issues.
This
study
introduces
an
AI-based
chatbot
that
allows
users
to
respond
validated
surveys
disorders
(PHQ-9,
GAD-7,
PCL-5)
in
a
natural,
conversation
manner
with
real-time
feedback.
The
study's
objectives
include
evaluating
chatbot's
usability
reducing
frequency
response
clarifications
while
accurately
interpreting
users'
responses.
system
was
assessed
running
hybrid
NLU
mode
(Phase
2;
N
=
587;
Mage
21.56,
SD
5.56,
67.8%
women)
after
being
trained
on
data
collected
rule-based
1;
274;
21.86,
5.50).
During
user-chatbot
interactions,
required
clarification
only
4.64%
time.
Using
AI
model,
could
understand
user
responses
85.65%
cases
interpret
similarly
human
annotators.
In
terms
usability,
perceived
as
more
engaging,
friendly,
easier
use
than
mode,
which
may
be
indirectly
attributed
enhanced
autonomy
provided
by
model.
Objective
With
the
increasing
global
burden
of
chronic
diseases,
there
is
potential
for
conversational
agents
(CAs)
to
assist
people
in
actively
managing
their
conditions.
This
paper
reviews
different
types
CAs
used
condition
management,
delving
into
characteristics
and
chosen
study
designs.
also
discusses
these
enhance
health
well-being
with
Methods
A
search
was
performed
February
2023
on
PubMed,
ACM
Digital
Library,
Scopus,
IEEE
Xplore.
Studies
were
included
if
they
focused
disease
management
or
prevention
systems
evaluated
target
user
groups.
Results
The
42
selected
studies
explored
diverse
across
11
Personalization
varied,
25
not
adapting
message
content,
while
others
incorporated
real-time
context.
Only
12
medical
records
conjunction
conditions
like
diabetes,
mental
health,
cardiovascular
issues,
cancer.
Despite
measurement
method
variations,
predominantly
emphasized
improved
outcomes
positive
attitudes
toward
CAs.
Conclusions
results
underscore
need
adapt
evolving
patient
needs,
customize
interventions,
incorporate
human
support
more
effective
care.
It
highlights
play
a
active
role
helping
individuals
manage
notes
value
linguistic
data
generated
during
interactions.
analysis
acknowledges
its
limitations
encourages
further
research
use
disease-specific
contexts.
Conversational
artificial
intelligence
(AI),
particularly
AI-based
conversational
agents
(CAs),
is
gaining
traction
in
mental
health
care.
Despite
their
growing
usage,
there
a
scarcity
of
comprehensive
evaluations
impact
on
and
well-being.
This
systematic
review
meta-analysis
aims
to
fill
this
gap
by
synthesizing
evidence
the
effectiveness
AI-
based
CAs
improving
factors
influencing
user
experience.
Twelve
databases
were
searched
for
experimental
studies
CAs’
effects
illnesses
psychological
well-being
published
before
May
26,
2023.
Out
7834
records,
35
eligible
identified
review,
out
which
15
randomized
controlled
trials
included
meta-analysis.The
revealed
that
significantly
reduce
symptoms
depression
(Hedge’s
g
0.64
[95%
CI
0.17-1.12])
distress
0.7
0.18-1.22]).
These
more
pronounced
are
multimodal,
generative
AI-based,
integrated
with
mobile/instant
messaging
apps,
targeting
clinical/subclinical
elderly
populations.
However,
CA-based
interventions
showed
no
significant
improvement
overall
0.32
-0.13
-
0.78]).
User
experience
was
largely
shaped
quality
human-AI
therapeutic
relationships,
content
engagement,
effective
communication.
findings
underscore
potential
addressing
issues.
Future
research
should
investigate
underlying
mechanisms
effectiveness,
assess
long-term
across
various
outcomes,
evaluate
safe
integration
large
language
models
(LLMs)
<p>The
concept
of
wellness,
as
proposed
by
Halbert
L.
Dunn,
recognizes
the
importance
multiple
dimensions,
such
social
and
mental
well-being,
in
maintaining
overall
health.
Neglecting
these
dimensions
can
have
long-term
negative
consequences
on
an
individual's
well-being.
In
context
traditional
in-person
therapy
sessions,
efforts
are
made
to
manually
identify
underlying
factors
that
contribute
disturbances,
factors,
if
triggered,
potentially
lead
severe
health
disorders.
Our
research
focuses
introducing
a
meticulous
task
aimed
at
identifying
indicators
wellness
detecting
their
presence
self-narrated
human
writings
Reddit
media
platform.
We
mentioned
Ethics
Broader
Impact.</p>
<p>The
concept
of
wellness,
as
proposed
by
Halbert
L.
Dunn,
recognizes
the
importance
multiple
dimensions,
such
social
and
mental
well-being,
in
maintaining
overall
health.
Neglecting
these
dimensions
can
have
long-term
negative
consequences
on
an
individual's
well-being.
In
context
traditional
in-person
therapy
sessions,
efforts
are
made
to
manually
identify
underlying
factors
that
contribute
disturbances,
factors,
if
triggered,
potentially
lead
severe
health
disorders.
Our
research
focuses
introducing
a
meticulous
task
aimed
at
identifying
indicators
wellness
detecting
their
presence
self-narrated
human
writings
Reddit
media
platform.
We
mentioned
Ethics
Broader
Impact.</p>
The
majority
of
individuals
presenting
with
mental
disorders
do
not
receive
health
services.
use
chatbots
for
is
now
frequently
discussed
as
a
means
to
increase
access
resources.
For
this
scoping
review
reviews
on
health,
we
performed
systematic
search
the
literature
Scopus,
Web
Science,
Pubmed
and
Dimensions.ai
identified
14
relevant
published
in
scientific
journals
which
publications
following
two
or
more
databases.
Three
additional
were
included
forward
backward
reference
examination.
17
are:
eight
(three
meta-analysis),
seven
unlabeled
reviews.
We
have
summarized
scope
well
their
findings.
Most
at
least
one
study
15
participants
less.
Few
report
an
important
proportion
RCTs
some
original
most
time
piloting
studies.
Overall,
examined
opinions
users
(generally
good
user’s
perceptions
although
find
deterrents),
features
chatbots,
outcomes
measures
studies
relying
effectiveness
(particularly
meta-analyses)
alleviation
disorder
symptomatology
meta-analyses
depression,
whereas
other
clinical
targets
might
need
further
research)
potential
assessment
health.
machine
learning
natural
language
processing
frameworks
(Dialogflow,
RASA)
seems
drive
increasing
number
conclude
that
there
but
progress
necessary
areas
affecting
interaction
dynamics:
reduction
frequent
lack
empathy
repetitiveness,
inclusion
user
within
chatbot
system,
memory
past
interactions.
Large
models
be
capable
reducing
part
these
issues
playing
increasingly
role
field.
Whether
data
privacy
can
guaranteed
also
question,
overlooked
