AI might reject tall structures, advise ‘greater square footage,’ predicts one expert Machine-learning still a long way to go before replacing human decision-making SAN DIEGO, USA: Twenty years ago when a major project was being conceived, it would be built and then the real estate representatives would take over, working to fill the colossal buildings. But today the physical completion of a development is just an initial part of the process. New technology is now being used to collect vast amounts of valuable data that can determine the success of both giga-projects and megaprojects. In some cases this will mean the downscaling of an initiative such as The Line, while others will be accelerated to meet strict deadlines such as the Saudi Expo 2030 project. In April 2024, Saudi Arabia’s Finance Minister Mohammed Al-Jadaan said some of the Kingdom’s projects would be adapted to current economic and geopolitical challenges. “I think that what we will see is a new rationalization of dates,” Naji Atallah, head of construction and manufacturing, EMEA Emerging Markets at Autodesk, told Arab News. Speaking on the sidelines of the recent Autodesk University 2024: The Design and Make Conference in San Diego, US, he said decisions would be based on new priorities. “Something like a new entertainment city won’t get the same prioritization as the Saudi Expo or projects for the World Cup, should Saudi Arabia win the bid, that have a definite deadline. But I think these projects will still be happening,” he added. On the reduction of Saudi Arabia’s The Line, Atallah said: “Being able or just having the vision of building something that big is the limit of our means … The new scale is more manageable.” But he said despite the change of size, it did not take away from the ambition. “The scale is smaller, but the ambition is still really large with what we are seeing now.” He said that projects like the Red Sea islands might not have the same scale of one aspect, but as a giga-project it is broken down into many different parts. “Part of the mandate of the Red Sea islands project is to create new tourism opportunities,” Atallah explained. “AI will help in the better decision-making to help set new targets that are not necessarily about immediate financial return.” And in the case of the Red Sea project, a key example was the introduction of scuba diving in the area. There is still a lot of work to be done on AI — it is only as good as the information it receives. But as the data banks continue to grow, the technology will learn and become more knowledgeable about future and existing projects. Change is most likely with giga-projects, said Autodesk CEO and president Andrew Anagnost, in an interview with Arab News. “I don’t think giga-projects are going to go away … but I do think that more often than not AI is going to advise against these projects.” Anagnost said he believed AI was more likely going to advise against massively tall structures and instead suggest “greater square footage.” AI would also possibly suggest different kinds of capabilities inside, such as sustainable energy generation, or multiple-use buildings that could serve as a home and workspace. “I think AI is definitely going to challenge some of these projects,” he added. But the capabilities of AI are only as great as the data it receives, and we remain a long way away from computers taking over the world. Also there remains a lot of mistrust in the collection of data, but the more information companies have, the more cost-effective and reliable their products become. Design and construction are not new concepts; humans have been creating tools, shelter, and the means to build for millions of years. However, there is surprisingly little information available on the physical structures that exist around the world. The US cloud-based security and management company noted that its clients in the architecture, engineering and construction industry have quadrupled their data storage from 0.9 terabytes in 2017 to 3.5 TB in 2021. But according to the investment banking firm FMI Corp., 95.5 percent of the data that is being gathered by AEC firms is not being used. And this is information that could tell governments, designers and architects of today and the future how to more successfully develop existing and new products — whether a chair, or new city. AI is very much dependent on the information that he described as 3D data and which is still lacking. “It’s kind of a paradox. We live in a 3D world, but the 3D data is scarce,” Ousama Lakhdar-Ghazal, director of trusted AI at Autodesk, explained. “When we look around and see all these 3D objects around us, we can easily picture them, but to have that represented in a digital world is actually very .” The collection of the data continues, but there remains a lot to be learned and that can only happen as the amount of data is gathered. This information can help with predicting the flow of a flood, or the fuel consumption of a new, existing or future building. “Like humans function better the more information they have, AI operates the same way, it needs to learn,” Lakhdar-Ghazal said. And the learnings of AI are entirely reliant on the information it is given, so it is still influenced by human input. “We are hoping that AI might at some point be able to maybe help solve some of the societal problems we face — that’s the driver,” he added. On concerns over the evolution of AI, Lakhdar-Ghazal acknowledged that society tends to fear what it does not know. “Most of the people working on AI are at a level (of a) Ph.D. (graduate), but for laypeople there’s a lot of unknown, there’s a lot of not understanding how it actually works.” While fear of technological advances is not new, he said it would take time to educate people to accept that the benefits outweigh potential drawbacks. “The point of AI is to help solve tangible problems. But it would still be up to humans to make the decision. AI can help identify labor-intensive, high-cost, low-return tasks, and help cut overheads.” But the practice of saving time has its own limitations and at some point optimization is reached — there is no more time that can be saved. But Lakhdar-Ghazal said the focus can always be shifted to improve areas including fuel efficiency, or other working practices, to cut overheads.