AI in construction: productivity gains and legal risks

Why AI matters to UK construction productivity

Productivity improvement is vital to the output and profitability of the construction sector. The RICS Construction Productivity Report of 9 March 2026 quote’s McKinsey’s 2024 research that global construction productivity grew by only 0.4% annually between 2000 and 2022 (a total of just 10% in 22 years) compared with about 2% every year for the total global economy. 

The RICS survey shows an improvement in the last 12 months but the net balance in construction reporting productivity enhancement in the UK is only 18%, compared with 22% in the Americas and 27% in Europe.

What can be done to improve it? The RICS survey highlights that the shortage of skilled labour is the only factor rated as high impact across all surveyed regions of the globe; 47% of responses rated upskilling the workforce as the best intervention to improve productivity. 

Digitalisation had more mixed support – only 17% in the UK. However, every construction expo that I attend these days focuses heavily on AI. It has to be part of the solution and surely the inevitability of adopting AI and other digitalisation is one of the reasons why the workforce needs upskilling.

Current adoption of AI across construction

The RICS artificial intelligence in construction report of September 2025 showed that 56% of surveyed investors planned to spend more on AI. However, 45% of responses reported no AI implementation at all in their organisation, with another 34% in an early pilot-stage. Only 12% reported regular use of AI in specific processes and a mere 1.5% across multiple processes.

A year is a long time in AI. There are clear signs of increasing adoption of AI in the architects’ profession. According to the RIBA AI Report 2026 :

• 74% of surveyed practices now use AI in at least some projects. The figure was only 41% in 2024;
• 73% report that AI has improved productivity; and
• 57% report a positive return on investment.

However, although 26% thought that AI has improved outcomes for clients or communities in relation to buildings, 29% disagreed, with the rest on the fence.

Interestingly, AI use is similar in architects with up to 30 years’ experience (at 80-85%) but trails off to 59% for those with more.

What is holding back investment in construction AI?

Perhaps the slow adoption of AI in construction is symptomatic of an industry that has for years had relatively slow growth in productivity. Perhaps the industry needs to be convinced of the long-term benefits of investing in a new technology that is still at an early stage of development, during a period of global shocks and economic uncertainty.

Other uncertainties include cost: will providers increase their charges as they look to recoup their start-up costs and their customers become increasingly reliant on AI? Will your provider still be around to supply and maintain the tool in a few years’ time? Will their product soon become obsolete?

One thing does seem to be clear, though: the influence of AI in construction is going to grow and become integral to the industry in years to come.

Issues to bear in mind when using AI in the construction industry

While many of these issues are not unique to construction and will be replicated across other industries, it is worth noting that in the case of construction, there is so far very little or nothing in standard form building contracts, professional appointments or collateral warranties that deals specifically with AI or other digital products. We expect this to change as AI grows in importance to the sector, and as lawyers as well as everyone else in the industry better understand how AI will be used, what risks need dealing with, who should best take on those risks and how to ensure adequate recourse when things go wrong.

It may be that the standard forms are already adequate to deal with some issues. It can also be the case that something labelled as AI isn’t a big departure from a tool that has an established track record in the industry. But these are not reasons for complacency – careful thought is needed where any new tool or system is to be used. 

AI companies may also impose unbalanced terms on their purchasers, and those risks need to be understood.

One thing is sure: when paying for an AI tool or contracting with someone who intends to use AI, you need to understand it well, so as to have a better idea of how to embrace it while mitigating risk.

A non-exhaustive list of other risks includes:

• Hallucination: The legal sector has seen too many instances of AI producing fictitious case references to back up legal arguments in court actions. Presumably similar risks apply in other fields. It is essential that experienced staff check the output of AI for this type of risk and others.
• Rubbish in, rubbish out: The AI can only be as good as the information it has. Is it coming up with a solution based on faulty information, or perhaps on good data but data which doesn’t deal with the specific circumstances of your project?
• Infringement of copyright: The AI may come up with a solution or design based on someone else’s intellectual property. What happens if you get sued for breach of their copyright? Will the AI provider pay up? Will your professional indemnity insurance (if any) help? If the AI solution was obtained by a professional or contractor instructed by you or by a subcontractor, are they good for your potential loss? Will their insurance back them up; if so, subject to what caps?
• Your own intellectual property: You need to guard against it being shared by staff with the AI tool.
• Will the AI fully comply with local law, such as building regulations?
• Might members or your staff use AI tools that you have not approved? This is just one risk that makes internal policies and training essential.
• Privacy, data protection and governance: For example, data protection law can require additional disclosures in privacy policies which may include the logic involved, but the AI provider may be unwilling to provide this information. Also, take care that staff don’t share information with the AI that is protected data. Data encryption and privacy-preserving AI are being developed – check for that.
• Fairness and bias: Has the AI tool been built so as to comply with local equality and diversity law and those of any other jurisdictions you deal with?  
• Contractual obligations: Have your business partners imposed on you any requirements that are relevant to your use of AI?
• Key personnel: If your professional team, contractor or subcontractor intend to use AI, are they dependent on the AI expertise of certain members of staff? If so, should you require that those persons are involved and replaced appropriately if they leave?
• Are there any relevant industry standards? Trade and professional bodies might be able to help identify these.
• Contingency planning: If your professional or contractor is using AI, do you need to obtain a licence to use the relevant software in case, for example, your business partners cease to exist?
• Could the AI go rogue or develop a new problem over time? Keep abreast of developments and keep checking the outputs.
• Does the AI tool enable you adequately to check the outputs? You may not know how it has reached its conclusions.

The potential for AI in Construction

Time will reveal in how many ways AI can help the construction industry, but these should, over time, include:

• Design of schemes – AI may help to incorporate best practice in placemaking
• Choosing the best procurement route
• Clash detection between different professionals’ and contractors’ designs
• Choice and drafting of contracts
• Interrogation of contract documents, for example in a dispute
• Production of contract documents from invitations to tender to programmes, specifications, bills of quantities and so on
• Monitoring of work, including health and safety
• Record keeping
• The Golden Thread of Information for higher risk buildings
• Digital twins and Building Information Modelling

Help with AI adoption and essential takeaways

The government has recently published assistance for government departments to help identify and manage risks associated with the development, deployment, procurement and use of AI solutions. There should be a lot in it to help the private sector too. AI Risk Management Toolkit: guidance - GOV.UK

Industry and professional bodies are all alive to the opportunities – but also the risks – of AI. The due diligence required for successful AI adoption is expansive, but a three-point checklist can help focus discussions and strategic adoption:

1. Get your lawyers to check your contracts for assistance with AI risks – both generally and with a particular eye on specific risks you identify in relation to a particular product or supplier; ideas on appropriate amendments are developing.
2. Tool up on in-house and external AI expertise.
3. Get your AI policy/policies in place and don’t treat them as a tick box exercise – treat them as an opportunity for the firm to consider the risks and how to mitigate them, and how to educate the team on the issues.

AI has the potential to help close construction’s productivity gap, but only if it is adopted with clear governance, informed contracts and proper human oversight. Businesses that engage with the technology now — while taking the legal, data and operational risks seriously — will be better placed to capture its benefits safely and commercially.