Age-group perspectives on large language models in the architecture, engineering, and construction industry: usage patterns, adoption expectations, trust, and future outlook

As large language models (LLMs) gain traction across industries, the architecture, engineering, and construction (AEC) industry is increasingly exploring their potential for documentation, analytics, and decision support. Successful adoption depends on how practitioners perceive and engage with these tools. This study provides an age-stratified empirical baseline of LLM usage, expectations, trust, and future outlook, based on a cross-sectional survey of 91 AEC professionals in three age groups (18–26, 27–42, 43–77). Younger and middle-aged respondents report significantly higher familiarity and more frequent use than older respondents, yet perceived usefulness among users is similar across groups. Expectations differ significantly for 11 of 12 assistant capabilities after false discovery rate correction, with younger professionals endorsing inspection and field-level support while all groups prioritize document drafting, summarization, and code querying. Trust is near the scale midpoint (slightly below neutral) across groups, with shared caution toward safety-critical applications consistent with calibrated reliance and lower perceived transparency among older respondents. Most respondents expect LLMs to become important in AEC practice within several years. The findings indicate that observed differences may reflect uneven exposure and skill development rather than dispositional resistance, and support a survey-informed practice framework that scales oversight from professional review to deterministic-system integration for safety-critical use.