Developers differ in personality traits that shape how they approach software development tasks, yet AI coding assistants either behave uniformly or offer personalization features whose fit with individual developers must be found through trial and error. This is because it remains unclear how personality shapes developers' experiences with AI assistants, and thus which traits personalization should account for. In other words, we provide tools with no underlying theory on what works when and why. We take a first step towards such a theory with a systematic scoping review on how users' personality relates to their perception of LLM-based tools. We identified twelve studies through general search engines, major Software Engineering (SE) venues, AI tooling, and snowballing. They all reported associations between personality and perceptions, particularly perceived usefulness, yet none addressed this in a SE context. Building on these findings, we conducted an exploratory two-wave survey with 29 and 11 developers, collecting first empirical evidence of how developers' personality (measured using the Big Five Model) relates to the perception of AI coding assistants. Higher Extraversion was associated with more positive private-use experiences in both waves, while Agreeableness was associated with negative professional experiences in the first wave and positive ones in the second. These results provide first indications that the personality effects found in the literature extend to developers. Although further studies with bigger samples are needed to build a theory, our work takes the first step towards an empirical grounded theory of developer personality and AI personalization traits.
Figures & tables
Fig. 1: Study Structure
ID
Aspect
Criterion
Inclusion
S1
Study type
Primary data source is human-centered study.
S2
Topic
Studies the impact of the user’s personality on their perception of LLMs; personality must appear as a research question, hypothesis, or other primary research objective.
S3
Technology
Focuses on LLMs, AI coding assistants, or general-purpose LLM-based tools; no specialized non-SE LLM-applications.
S4
Context
Any study context.
S5
Personality measure
Assesses personality with an established multidimensional personality framework.
S6
Availability
Publicly available in English.
TABLE I: Inclusion and exclusion criteria.
Fig. 2: Scoping review structure and intermediate results
Framework
Outcome Measures
O
C
E
A
N
Other
[ 31 ]
BFI-2 [ 36 ]
Usefulness, likeability, follow intention, creepiness of advice
n.s.
n.s.
↓ PU
↑ PU ↑ Likeability ↑ Follow intention
↓ PU ↓ Advice from friend rating
[ 32 ]
BFI [ 59 ] (partial)
TAM (PEOU, PU, CI)
↑ PU, PEOU ↑ CI (indirect)
↑ PEOU (limited)
n.s.
[ 60 ]
BFI-44 [ 35 ] (partial)
Tech. acceptance, LLM usage, cognitive dissonance
n.s.
↓ Acceptance ↑ Dissonance (small, indirect)
[ 61 ]
BFI-44 [ 35 ]
TAM (PU, PEOU, AT, BI)
↑ PU
↑ PU, PEOU
↑ PEOU
↑ PU
↓ PEOU
[ 62 ]
CP5F [ 63 ]
Cognitive, affective, behavioral attitude
n.s.
n.s.
↑ Open attitude ↑ Behavior
n.s.
n.s.
[ 64 ]
IPIP-50 [ 65 ]
GenAI use and perceptions (frequency, purpose, usefulness, preferences)
↑ Preference of traditional tools
↓ Usage frequency
↑ Preference of GenAI
↑ Ethical concerns ↓ Performance
↑ Learning impact ↓ Concerns
TABLE II: Overview of review stuies. Key (not all) findings organized by personality dimensions. Empty trait columns indicate a trait not evaluated on its own. Arrows indicate how a higher presence of a trait correlates with higher ( ↑ ) or lower ( ↓ ) outcome values.
Avg
Δ
α
CIL
CIU
Construct
S1
S2
S1
S2
S1
S2
S1
S2
experpriv
4.1
4.4
0.40
0.68
0.77
0.46
0.49
0.83
0.93
expect neg
2.8
0.71
0.51
0.85
exper neg
2.6
2.6
0.10
0.51
0.55
0.17
-0.08
0.75
0.86
expect pos
3.5
0.8
0.66
0.89
exper pos
3.3
3.7
0.50
0.79
0.58
0.64
0.02
0.89
0.87
TABLE III: Internal consistency and averages of constructs in Survey 1 (S1) and Survey 2 (S2): total mean score (Avg), mean change per person between S1 and S2 ( Δ ), Cronbach’s alpha ( α ), and 95% confidence interval bounds ( CIL , CIU ).
Coefficient estimates
Model fit
Coef.
SE
t
P>|t|
R 2
RA2
F
pF
pJB
pBP
Survey 1 (N=29)
positive expectations
c
2.45
1.91
1.28
0.21
0.05
-0.15
0.27
0.93
0.23
0.17
O
0.12
0.30
0.40
0.70
C
-0.00
0.23
-0.00
1.00
E
0.18
0.21
0.82
0.42
TABLE IV: Regression results per survey wave: coefficients (Coef.), standard errors ( SE ), t-statistics (t) and p-values (P>|t|). Model fit: (adjusted) R2 ( RA2 ), F-statistic and p-value ( pF ), and Jarque–Bera ( pJB ) and Breusch–Pagan ( pBP ) test p-values.
Fig. 3: Pearson correlations (r) between individual construct items and personality dimensions that significantly predicted the corresponding construct
School of Education University of California Irvine, CA, USA · School of Education & Department of Cognitive Science University of California Irvine, CA, USA