Can artificial intelligence help solve the housing crisis in the Gulf region?

(TNS) – Can artificial intelligence design a viable building? If so, how quickly will homes be built in the Bay Area?

In a housing-starved market like the Bay Area, some real estate developers are turning to the promise of artificial intelligence, hoping to reduce design and construction time and save money in the process. But in a region known for allowing nightmares — especially in San Francisco, where housing construction takes nearly a year longer than anywhere else in the state — all that computational brainpower can find itself wrapped up in red tape.

It promises a suite of cutting-edge products using cutting-edge generative AI — think ChatGPT, but in 3D — to automatically run code-compliant wiring through digital buildings, visualize imaginary structures from a sketch, and schedule contractors to the minute.

A new artificial intelligence tool from San Francisco software giant Autodesk can produce thousands of building designs in hours instead of days or weeks, taking into account 10 criteria ranging from cost, carbon footprint and ease of living.

The backup of Interstate 880 southeast of Jack London Square is among the first test cases of this technology moving from the computer screen to the real world.

Once home to the Phoenix Iron Works, the site could eventually host 316 housing units, about a third of which are either low-income or supportive housing, as well as thousands of square feet of office and light industrial space.

Working with studio, one- and two-bedroom units made by Factory_OS, a modular housing manufacturer in Vallejo, the designers used artificial intelligence software from Autodesk, whose computer-aided design software is already the industry standard, to quickly design thousands of configurations. Buildings can fit on the plot of land, and how much carbon will be emitted to get them there.

“Architecture and buildings account for about 40% of the global carbon problem,” said David Benjamin, director of architecture, engineering and design research at Autodesk, whose team applied the software to the project.

“The impossible problem that AI can perhaps help solve is how will we significantly increase the total floor area while dramatically reducing the total carbon emissions of all buildings?” Benjamin said.

Using the parameters of the former Phoenix site and the fixed dimensions of the Factory_OS modules, Benjamin’s team modeled how seven smaller structures and a large ribbon-shaped building could be moved to improve green space, walkability to BART, sunlight, noise abatement and other criteria.

Ultimately, two designs — one with a large central green area and the other with a more distributed landscape with smaller buildings scattered in a checkerboard arrangement — went before the Oakland Planning Commission. The first plan, called “Central Park,” was eventually chosen, Benjamin said.

Using the software means “we know exactly what we’re building ahead of time,” Jamie Hettishaw, development director at Holliday Development, the real estate development arm of Factory_OS, and project lead told Phoenix. That way, “we’re not dealing with a lot of iterations of site plans,” moving back and forth between stakeholders, he said.

The first phase of the project, which has been authorized and started, includes about 100 units, 51 of which will go to formerly homeless people. Hiteshew estimated that Autodesk cut about six months out of the design process.

However, real estate developers have warned that AI will only solve part of their problems. Planning and design typically represents only about 5% of the total project costs for a Bay Area development, Carolyn Bookhart, director of real estate development at the nonprofit Community Development Resources, said in an email. Other estimates range as high as 10%.

Bookhart, speaking generally rather than specifically about the Phoenix project, said 60% to 70% of project costs typically come from the actual construction.

Ryan McNulty, principal architect at MBH Architects, which also designed the new development using Autodesk software, said the process was faster than having to repeatedly redraw designs to focus on a particular factor, such as carbon emissions or cost. project.

A human architect can optimize certain features, such as heating or privacy, while the system “can explore hundreds of options very quickly,” Benjamin said.

The tool shaved hours off design time that would have otherwise been done manually, McNulty said, adding: “For us, it’s a tool to understand what we’re trying to do and make decisions faster and more effectively.” But he warned that it is “not a silver bullet that allows us to press a button and build a building.”

Autodesk isn’t the only software maker or researcher using AI in design and construction. Companies including Riveia, HD Lab and TestFit are also testing variations of so-called generative design that allows creatives to imagine buildings inside and out.

A project at non-profit technology research institute SRI International has created generative AI software to turn drawings into potential architectural designs and more efficiently map onto existing buildings, in partnership with Japanese construction giant Obayashi Corp.

Eric Yeh, a senior computer scientist at SRI, said his team is working on ways to change aspects of AI-generated building design using text prompts. He hopes to come up with tools that can map out the initial design of a building, including its cost and other features, in the near future.

Another company called Augmenta, founded by Autodesk alum Francesco Iorio, aims to take 3D models of buildings and use AI to automatically determine the best placement of interior systems like electrical wiring, cost calculations, electrical codes, and parts availability.

Automating parts of the building design phase could also represent huge savings in large, non-residential projects such as hospitals or Offices.

Software company Slate in Pleasanton provides construction managers with a single database that captures all the deadlines and moving parts of a project, from subcontractors to open requests for information. It even represents weather conditions. CEO Trevor Schick said he hopes to build a software tool by the end of this year that can generate designs based on a range of criteria, along with a detailed project and construction schedule.

From this perspective, AI programs appear to be transformative. However, for housing developers facing slow-moving financing and permitting processes, this may only represent an incremental improvement.

Case in point: The final development permit to begin construction on the Phoenix site was approved by the Oakland Planning Commission nearly five years ago, according to an email from Jane Walsh, Oakland’s public information officer. She said planning approvals had been extended twice to give developers more time.

One nonprofit housing expert has long been skeptical about how much impact artificial intelligence could have on affordable housing development.

“The nonprofit housing industry is already without a doubt the most skilled manager in construction development,” said Sam Moss, executive director of the San Francisco Housing Development Corporation, speaking generally rather than about the Phoenix project.

Moss acknowledged that there are opportunities to make the process more efficient. “But in terms of making a massive impact where we’re going to deliver all these magic units a year early, it’s more important to focus on dramatically increasing our funding” and passing pro-housing legislation at the state level, he said.

In the Phoenix project, construction is progressing well. But the project has not yet received a final development permit for the rest of its more than 200 market-rate units, according to Walsh, Oakland’s public information officer.

Said Hiteshew, developer. “We do not have specific prices for the rest of Phase 2 or the entire project.”

©2023 San Francisco Chronicle, Distributed by Tribune Content Agency, LLC.

You may also like...

Leave a Reply

Your email address will not be published. Required fields are marked *