The recent Digital Construction Summit in Dubai brought together industry leaders to discuss what many consider to be the most transformative technology facing our sector: artificial intelligence. As moderator of a panel discussion with experienced construction technology experts, I witnessed an interesting tension emerge. This should be both a warning and an opportunity for the construction industry in the Middle East.
The uncomfortable truth about our digital maturity
Digital transformation efforts in the construction industry span more than 15 years, but the realities on many sites remain harsh. Companies continue to rely on Excel spreadsheets and WhatsApp for field management instead of implementing proper centralized and common data environment solutions.
This fragmentation has real consequences. In my law practice, I frequently encounter disputes where strong cases are undermined by insufficient contemporaneous field reporting, recording, and documentation, even though digital recording has never been easier. Experts on the panel highlighted that a staggering 80% of data is lost during the handoff from design to construction to operations phases due to stakeholder fragmentation and inadequate quality processes.
Lessons learned from the implementation of digital tools such as Building Information Modeling (BIM) are instructive. The industry has repeatedly introduced technology without proper processes, resulting in data being accumulated without a clear usage strategy. We cannot afford to repeat this mistake with AI.
AI rush: demand without understanding
Panel members revealed a strange contradiction. Clients specifically request AI implementation into their projects without understanding the application, often defaulting to a narrow understanding of generative AI and missing out on the potential for broader automation.
This knowledge gap is being exploited. Over-promising by consultants has led to unrealistic expectations of current AI capabilities, leading to implementation failures and client dissatisfaction. The construction industry is already cautious about adopting the technology, but there is a risk of further resistance if early AI implementations fail to live up to their exaggerated promises.
What makes AI uniquely difficult is its “black box” nature. These systems require unprecedented trust without explainable processes, a particularly difficult proposition for an industry that has historically had a slow pace of technology adoption.
Governance gap: Who takes the risk?
The legal implications of AI implementation in the construction sector are profound but largely unaddressed. When AI systems serve multiple project participants simultaneously, significant questions arise regarding the allocation of responsibility and accuracy standards. Current government regulations focus on general data sharing and cybersecurity and do not address construction-specific AI applications or responsibility allocation.
Our panel identified several governance imperatives.
- Intellectual property protection: Employees can accidentally expose sensitive company information through open source AI platforms without proper training. Robust IP protection protocols are no longer optional.
- human surveillance: The recommended approach is a human-involved methodology, rather than full automation, that breaks down complex tasks such as infrastructure design into reviewable stages, with expert validation at each checkpoint. This requires careful definition of acceptable AI applications and limited use cases. For example, embracing QA/QC automation with human oversight while avoiding AI-only safety oversight.
- Multi-stakeholder responsibility: As AI tools become incorporated into the delivery of collaborative projects, contractual frameworks will need to evolve to address shared responsibility and risk allocation in ways that current standard formats simply do not allow for.
Warning from technology leaders
Perhaps most revealing was the data shared from the technology sector itself. An internal study at one construction software solutions company found that while 90% of developers deployed AI tools to significantly improve documentation efficiency, certain tools, such as Copilot, actually reduced productivity by 20% due to falsified results. If technology experts, the most digitally savvy users, experience such mixed results, what does this mean for construction site implementation?
The Middle East advantage: innovation through experimentation
Despite these challenges, my panel of construction technology experts and I believe there are real opportunities for the Middle East’s construction sector. Europe’s regulatory framework can create a competitive disadvantage for companies in China and the United States that operate in less restrictive environments, positioning the UAE as an experimental market for innovation.
Compared to highly regulated markets, the region’s experimental approach to technology adoption could create a competitive position for construction companies in the Middle East, but only if they approach AI adoption strategically.
Roadmap for responsible AI adoption
Based on the panel’s collective expertise and insights, as well as my legal knowledge and experience in the region, I recommend construction companies in the Middle East consider the following framework.
1. Basics come first: Avoid technology jumps without proper data infrastructure. Before moving forward with new AI implementations, make sure you understand the fundamentals of BIM and data quality. AI cannot fix bad data. It will only exacerbate existing problems.
2. Phased rollout: Start with controlled experiments within a single organization before expanding to a multi-stakeholder environment where accountability and accuracy become legally complex. Learn your AI tools in a controlled environment before scaling up and out.
3. Invest in human resources: Manager training and mindset changes remain important as the construction industry’s traditional resistance to new technology is amplified by the trust-demanding nature of AI. Technology is only as effective as the people who use it.
4. Augment, not replace.: Follow gradual adoption steps to position AI as a tool that enhances traditional methods rather than completely transforming established processes, while maintaining the integration of human expertise. The goal (at least for now) should be data complementation, not outright replacement of proven practices.
5. Build clear governance: Establish internal policies around acceptable AI use, intellectual property protection, data security, and allocation of liability before a dispute arises. Contractual frameworks need to evolve with technology deployments.
It’s important to do it right
The construction industry is at a crossroads. We can tackle AI implementation head-on and repeat the mistakes of previous technology implementations, or we can take a measured and strategic approach built on strong digital foundations.
The Middle East has a unique opportunity to take the lead in this space, leveraging the region’s openness to innovation while learning from the mistakes of early adopters in other regions. But this requires honesty about current digital maturity, realistic expectations about AI capabilities, and a robust governance framework to manage risk.
The question is not whether AI will transform construction, but it will. The question is: Can you manage that transformation intelligently, or will you end up implementing technology without the processes, training, and governance required to realize its full value?
Karie Akeelah is a partner in Trowers & Hamlins’ international construction disputes team.
