← CHI 2026 map

CHI '26 · Honorable mention · full-paper review · confidence low

Exploring Design Practice with Generative AI: Perspectives from AEC Design Professionals

Yue Xu , Yi Wang , Weiyue Gao , Yichen Chai , Henry Been-Lirn Duh

This bundle is evidence-poor because the provided spans are only navigation/classification text, so the review cannot verify the paper’s claims from the body. At the metadata level, though, the paper looks like a useful mixed-methods descriptive study of GenAI use in AEC practice, with likely value in documenting adoption patterns and disclosure tensions.


Axes Lens

Rare contribution shape, typical evidence profile. The point here is not a score. It is to show what kind of claim the paper makes, and whether the evidence pattern is unusual or baseline in this 268 -review set.

Contribution shape

Knowledge form
descriptive knowledge typical · 92/268
Novelty type
empirical finding typical · 68/268
Abstraction level
practice typical · 85/268
Generalization target
organizational context typical · 20/268
Validation mode
mixed methods typical · 136/268

Evidence profile

Evidence strength
weak less common · 5/268
Claim alignment
weak less common · 5/268
Overclaim risk
high less common · 5/268

Review Summary

From the evidence actually provided, this review has to stay very conservative. The quoted spans are navigation or classification fragments such as “skip to main content” and “ACM Computing Classification System,” which do not support claims about the paper’s contribution, method, findings, or limitations. The abstract metadata does indicate that the paper is a mixed-methods study of 20 interviews and a survey of 191 professionals, and it frames the work as an empirical study of GenAI adoption in AEC design practice. That is enough to infer the likely genre of contribution at a high level, but not enough to verify the paper’s actual novelty or the strength of its evidence. In particular, the packet does not include the results, discussion, or limitations sections, so there is no grounded basis for saying what is surprising, what is new relative to prior CHI or AEC work, or where the claims are bounded. Because of that, the safest expert reading is that the paper may be positioned as a descriptive empirical contribution about professional practice, but the supplied evidence is insufficient to evaluate whether it truly offers a first-of-its-kind account, whether the mixed-methods design is executed rigorously, or whether the conclusions are appropriately scoped. The review therefore flags high overclaim risk and weak claim-evidence alignment, not because the paper is necessarily weak, but because the evidence packet is incomplete for a full CHI review. The main limitation is not a flaw in the paper itself but a flaw in the available review substrate: without body text, the review cannot substantiate novelty, validation scope, or limitations with exact quotations and offsets. That means the bundle should be treated as provisional and manually checked before publication.

What Changed

Canon before

Prior CHI work on GenAI in professional design practice is typically fragmented, often focusing on general creative work or isolated tool use rather than AEC-specific empirical adoption patterns.

Departure from common sense

The paper’s core contribution is not a surprising mechanism but a grounded account of how AEC professionals actually use GenAI in practice. That matters because it replaces generic assumptions about “AI in design” with domain-specific evidence about conceptual design, collaboration, and disclosure norms.

Actual novelty

The novelty is an empirical, AEC-specific account of GenAI adoption built from mixed methods rather than a new algorithm or interface. The paper combines interviews and survey evidence to characterize where GenAI is used, what value it provides, and how professionals negotiate tensions around creativity, unpredictability, and public disclosure. This is a descriptive contribution that extends CHI’s understanding of professional GenAI use into a domain with distinct constraints and workflows.

Evidence

The supplied evidence is limited to front matter and ACM navigation/classification text, so the bundle cannot directly quote the paper’s findings, methods, or limitations from the body. The abstract metadata, however, indicates a mixed-methods study with 20 interviews and a survey of 191 professionals, and frames the contribution as an empirical account of GenAI use in AEC design practice.

“ACM Computing Classification System”

actual novelty · ACM Computing Classification System · confidence 0.01

“skip to main content”

departure from common sense · Front Matter · confidence 0.01

“ We use cookies to ensure that we give you the best experience on our website.”

limitation · Advanced Search · confidence 0.01

“ Advanced Search Journals Magazines Proceedings Books SIGs Conferences Institutions People More”

validation scope · Register · confidence 0.01

Limits

Method limits

No method or results section text was available in the provided evidence spans, so the review cannot verify sampling, analysis procedures, or measurement quality from the paper body. Any assessment of rigor must therefore remain provisional and based only on abstract-level metadata.

Deployment limits

No deployment or implementation discussion was available in the provided evidence spans, so limits on real-world adoption, organizational rollout, or operational integration cannot be grounded from the supplied text.

Boundary conditions

The supplied evidence does not expose the paper’s substantive body text, so boundary conditions such as project type, firm context, or task phase cannot be verified. The abstract suggests the findings are especially relevant to conceptual design and collaboration-heavy AEC settings, but that remains metadata-level inference.

Position in field

Based on the abstract metadata, the paper appears positioned as an empirical CHI study of GenAI adoption in AEC design practice, aiming to fill a gap in professional adoption evidence. The contribution is best read as field-building descriptive evidence rather than a new interaction technique or system.

Abstract