RIBA AI Report 2026 charts the growing role of artificial intelligence in architectural practice
August 2026Artificial intelligence (AI) is becoming an increasingly established part of architectural practice. As we noted in our previous article regarding the RIBA AI Report 2025, last year marked a shift from cautious experimentation towards routine use. The RIBA AI Report 2026 (the Report) shows that this transition has continued at pace. AI is now being used more widely, more intensively and across a broader range of activities. At the same time, the Report reflects a more mature appreciation of the legal, professional and commercial risks that follow.
Findings of the Report
The headline figure is striking. In 2024, 41% of practices reported using AI at least occasionally. That increased to 59% in 2025 and has now reached 74%. Large practices continue to lead adoption, with more than 90% of practices with 50 or more staff using AI on at least some projects, compared with 84% of medium-sized and 63% of smaller practices.
AI is also moving beyond image generation. Early-stage design visualisation remains the most common use, reported by 56% of AI users, but 40% now use AI for standards and regulatory compliance checking, 32% for practice management and 24% for specification writing or technical design.
The productivity case is becoming clearer. Among respondents using AI, 73% report improved productivity and 57% of those assessing return on investment report a positive return. However, only 17% agree that their designs are “better” because of AI. The evidence is therefore stronger that AI can make practices faster and more efficient than that it improves the quality of the final design.
Notes of caution
The profession’s concerns have also sharpened. More than three-quarters of respondents (76%) agree that AI increases the risk of imitation, raising familiar issues around intellectual property, authorship and originality. A majority (55%) believe AI will enable people without sufficient professional qualifications to design buildings, creating concerns around competence, building safety and accountability.
RIBA also highlights the risks of hallucination and unreliable outputs, data security, privacy and professional insurance. Architects who participated in RIBA’s research raised concerns about whether professional indemnity insurance would respond to AI-related errors and the potential impact of AI use on insurance premiums. These issues reinforce the point made in our previous article regarding AI readiness across the professions in that, regardless of the technology used, accountability for professional work remains with the professional.
Human oversight is therefore critical. The Report repeatedly emphasises that AI is not a replacement for professional judgement, particularly where it is used for compliance checking, technical design or other safety-critical tasks.
Governance and AI policies
Perhaps unsurprisingly given the pace of change, there remains a significant gap between adoption and governance. Only 19% of practices currently have a documented AI policy, although a further 17% are developing one and 15% plan to do so. Larger practices are considerably further ahead: 56% already have an AI policy.
This is particularly significant as AI moves from informal experimentation into live project work. As discussed in our previous articles, organisations should have clear controls around approved tools, confidentiality and data handling, checking and verification of outputs, intellectual property, record keeping and responsibility for final decisions.
RIBA itself is moving towards a more structured framework. During 2026 it intends to develop an AI Overlay to the RIBA Plan of Work and a policy for the responsible use of AI, outlining best practice.
The future of the profession
The Report also raises a longer-term concern: the effect of automation on professional development. Some 61% of respondents believe AI will make it harder for early-career professionals to gain the skills they need, because routine tasks susceptible to automation are often the same tasks through which junior architects build experience.
Yet the profession remains confident in the continuing importance of human input: 77% agree that AI can never replace human creativity. The Report therefore points towards augmentation rather than substitution, but only where practices retain meaningful human oversight and ensure that efficiency does not come at the expense of competence.
Key points to consider
The RIBA AI Report 2026 confirms that AI adoption is no longer a future issue for architectural practices. The question is increasingly how AI is governed, checked and integrated into professional work.
The legal risks identified in our 2025 review remain, but they are becoming more immediate as AI is used on live projects and in higher-risk activities. Clear AI policies, appropriate training, robust checking processes and careful consideration of insurance, intellectual property and data security will be central to managing those risks.
Where AI is being used by architects and/or their subconsultants it is also increasingly important to ensure this does not conflict with the appointment and to consider including provisions in the appointment and sub-consultancy agreement to reflect and manage the risks of AI.
If you would like to discuss this article or how we can assist in relation to the legal risks associated with AI and/or an AI policy, please contact Andrew Croft or Anna Benz.
Further Reading
Download PDF
