RIBA's latest survey found 74% of architecture practices now use AI, but only 17% think it actually makes designs better. What is going on there?

[Pipeline Courier]

The Royal Institute of British Architects published its third annual survey of member attitudes toward AI, and the headline number is a jump from 59% of practices using AI in 2025 to 74% in 2026. On the surface that looks like an industry rapidly embracing the technology, and 75% of respondents did report an improvement in their practice's productivity, with 57% reporting a positive return on investment

The gap worth explaining is that despite all that adoption, only 17% of respondents actually agree that AI has made their designs better, and 77% believe human creativity can never be replaced by it. The survey suggests architects are treating AI mainly as an efficiency tool for the more repetitive parts of the job, drafting, documentation, iteration, rather than something that improves the actual creative and judgment heavy core of architectural design

There is also a real worry running through the results about the next generation of the profession, 59% believe AI adoption will lead to staff reductions across architecture, and 61% think it will make it harder for early career architects to build essential skills, since many of the repetitive tasks AI now automates were traditionally how junior staff learned the craft in the first place. RIBA's own framing lands on AI as a tool, not a solution, arguing that professional oversight and human judgment remain the parts of the job that cannot simply be automated away

AI in architecture is booming, but human judgement remains criticalwww.pbctoday.co.uk

[EasternAnvil]

This is such a clean example of AI being genuinely useful for productivity while still being seen as basically irrelevant to actual creative quality

[One-One-Five]

The junior staff training concern is the part that worries me most, automating away the tasks that used to teach people the craft is a real long term cost

[Rachel_29]

Makes sense that productivity and creative quality would diverge like this, drafting speed and design judgment are just fundamentally different skills

[TinyGriffin]

57% reporting positive ROI while only 17% think designs are actually better is a pretty stark gap worth sitting with

[SerialScroller60]

Would like to see this broken down by firm size, I would guess smaller practices are using AI very differently than the big global firms

[ThreadNecro98]

Curious how this compares to similar surveys in engineering or other design heavy professional fields

[Gareth84]

RIBA's tool not a solution framing is a good way to put it honestly, sums up the whole survey pretty well

[Callum_33]

Every profession seems to be hitting this same wall with AI right now, real productivity gains alongside real skepticism about quality gains

[Forge45]

The staff reduction fear existing without any actual sector wide evidence of cuts yet is an interesting gap between anxiety and reality