Brunei and Generative AI
Gender and Readiness
Similar adoption does not mean similar reported readiness.
Following the main Brunei and Generative AI report, this companion takes a closer look at one pattern in the 2025 survey data: female and male respondents were adopting Generative AI at similar rates, but described their understanding, confidence and certainty differently.
Start with adoption
At first, the difference looks small.
Overall Generative AI adoption was already high for both groups in this sample.
used GenAI frequently or occasionally
used GenAI frequently or occasionally
Base: main eligible sample. Female n=250; male n=181.
Among GenAI users, work-related use was almost identical for female and male respondents.
If access and use were the whole story, we might stop here.
But understanding and confidence were a different story.After adoption
The gap becomes much more visible.
Reported clear understanding of the difference between AI and Generative AI
Base: main eligible sample. Female n=250; male n=181.
Reported high confidence using Generative AI
Base: GenAI users. Female n=230; male n=164. High confidence = ratings 4–5 on a five-point scale.
These are self-reported measures. The survey did not test actual AI knowledge or competency. The findings should not be read as proof that one gender has greater or lesser AI ability.
Frequency matters, but does not remove the pattern
Even among people using GenAI more often, the difference remains.
Frequent users
Occasional users
Frequent use is part of the readiness story. In this sample, however, it does not by itself explain the gender pattern.
Reported uncertainty
Another pattern appears in the answers people were less certain about.
“Can Generative AI reflect Asian or Bruneian perspectives?”
Female respondents were also more likely to report uncertainty about whether they had encountered AI bias, whether they were aware of ethical AI guidelines and whether AI would replace jobs in their industry.
It may reflect limited exposure or confidence, but it can also reflect genuine ambiguity, caution or reluctance to make a definitive claim without enough evidence.
These differences should not be interpreted as one group being more ethical. Reliability is a perception measure, not a technical accuracy test.
Could something else explain it?
The analysis checked several obvious possibilities.
The understanding difference appears across the major education levels in this sample.
Clear understanding. Some subgroup sizes are modest; this does not prove a uniform gender effect within every education category.
Industry profiles differed. Male respondents were more represented in Technology & IT. But excluding Technology & IT respondents did not remove the headline pattern.
Exploratory logistic regression accounted for age, educational attainment, frequent use and whether respondents selected Technology & IT or Education & Training.
95% CI 1.74–4.46
95% CI 1.10–2.80
Exploratory models only. They do not establish that gender causes the differences and cannot account for unmeasured factors such as role, seniority, prior technical experience, training quality or workplace culture.
A closer look at Education & Training
Similar adoption. Similar reported training. Very different self-reported readiness.
Education & Training was the largest industry grouping in the sample, with 180 respondents.
Base: respondents selecting Education & Training. Female n=123; male n=57. *Confidence excludes non-users.
The survey does not identify comparable roles, subjects or seniority within this subgroup. A purpose-designed education study would be needed before generalising to educators more broadly.
Training access
A simple access explanation does not fit the data either.
*Employed/self-employed branch: female n=200; male n=141. Training demand is strong in both groups, while reported provision is similarly limited.
Equal availability may not mean equal learning experience. Relevance, opportunities to practise, peer support, encouragement, prior exposure and role expectations may all matter, but the survey did not measure these dimensions directly.
What should we take from this?
This report identifies the difference.
It does not yet explain it.
People adopting the same technology at similar rates can describe their readiness very differently.
Whether the difference reflects actual knowledge, self-assessment, workplace experience, role, culture, learning opportunities or other factors.
About this analysis
Read the findings with the right boundaries.
Methodology and important limitations
- This was an independent online sample and is not statistically representative of the Brunei population.
- Understanding and confidence are self-reported. No objective AI competency or knowledge test was administered.
- The survey gender item offered Female and Male response options only, so this analysis cannot represent gender identities outside those categories.
- The cross-sectional survey can identify associations but cannot establish why the differences exist or whether gender causes them.
- Adjusted analysis accounts for some broad factors, but not job function, seniority, prior technical experience, training quality or workplace culture.
- Some subgroup analyses have modest sample sizes.
- The data were collected in 2025. This is a secondary analysis of that survey period, not a current 2026 national measurement.
Read the full report
Brunei and Generative AI:
Gender and Readiness
The full report includes the wider analysis, statistical checks, interpretation and limitations behind the findings shown here.
Download the free reportPublished independently by Syafiq Bakar under Mr Syaf Insights · September 2026