At first glance, US real estate firm Opendoor offers a pretty good deal. People looking to sell their home can upload their address, basic details and some photos to the website. It scours the information, cross-references market data and makes an offer. If the offer is accepted, the seller is sped through the transaction process, chain and broker-free. Once sold, Opendoor makes some renovations and resells the home at a profit: a speedy \u2018flip\u2019 that cuts through the normally slow real estate system.<\/p>\n
The reality is not so straightforward. Buyers and sellers have complained bitterly about Opendoor: \u2018After a year living in it, the walls in the showers have all caved in and there is rotting wood behind the tile,\u2019 complains one typical online review. \u2018They used the cheapest contractors available.\u2019 \u2018The offer was a real low ball to start,\u2019 reports another. It\u2019s an age-old story of a company buying for less than market value, refurbishing and selling at a mark-up. But the difference is that Opendoor is powered by \u2018deep learning\u2019 algorithms that can do this job extra quickly and even more ruthlessly.\u00a0<\/span><\/p>\n \u2018iBuyers\u2019 like Opendoor, which has reportedly sold nearly 2 million US homes since 2016, show just how far AI tools have penetrated real estate. The site is unusual in being consumer-facing, rather than business-facing, and is particularly ambitious in that it attempts to value, buy and flip homes in one go. Opendoor has struggled to turn a consistent profit. But it is otherwise typical of current real estate AI: a tool that supercharges the treatment of buildings as assets to be traded and gambled.<\/p>\n Most commonly, real estate AI is used for investment analytics. Companies like Reonomy, Skyline AI and Cherre offer powerful investor data-crunching tools. Compass is an example of a \u2018lead generation\u2019 or \u2018real estate farming\u2019 AI software that helps real estate agents find people whose online activity suggests they are house-hunting. US-based Deepblocks pinpoints development opportunities based on migration, market and planning information. All mine data to maximise potential profits from homes and buildings.<\/p>\n There\u2019s also an emerging world of real estate AI that replaces human interactions. According to McKinsey: \u2018Simple requests from tenants, such as for routine maintenance, can prompt the [AI] copilot to directly contact a building\u2019s maintenance staff […] For high-stakes moments \u2013 such as a commercial lease negotiation with an office, warehouse, or retail tenant \u2013 a gen AI tool can [\u2026] craft a negotiation transcript.\u2019 There is some potential for AI to cut through bureaucracy here.<\/p>\n But there are also gaping digital inclusivity and surveillance risks, as well as question marks over how AI will handle the racial and socio-economic inequalities entrenched in the housing system. While some are optimistic about its ability to eliminate discrimination, studies have already found evidence of AI amplifying inequalities in mortgage approvals, for example.<\/p>\n Other softwares, often called Automated Valuation Models (AVMs), are dedicated to costing homes. These range from crunching local price index numbers and sales data to full-blown image scanning. Homer, from south-east Asian tech start-up Oh My Home, for example, visually assesses images of homes to check for structural damage as well as \u2018clutter, untidy rooms and outdated interiors\u2019. AVMs seek to reduce homes to single data points, metricising the messy balance of factors that make a space liveable and loveable into a single estimate.<\/p>\n \u2018Imagine a dystopian future where everyone repaints their house a specific colour to impress the Google Street View car\u2019<\/p><\/blockquote>\n This is perhaps where AI most immediately threatens the art of the architect. \u2018Imagine a dystopian future,\u2019 warn researchers at MIT, \u2018where everyone repaints their house a specific colour to game the system and impress the Google Street View car.\u2019<\/p>\n While none of the above tackle the most pressing issues in housing, such as environmental impacts, affordability or homelessness, there is potential for more inspiring approaches. The Turing Institute thinks AI could help planners navigate the technicalities of planning applications and spend more time on political considerations, and some AI tools help increase community engagement. Non-profit JustFix builds digital tools that help tenants exercise their rights. Predictive maintenance tools, materials passports and open source BIM emerging from engineering firms could help democratise our knowledge of our cities and infrastructure.<\/p>\n Flashy new AI design tools and chatbots may be distracting us from how quickly and deeply AI is penetrating the built environment. Investors have found a bureaucracy-laden world ripe for \u2018disruption\u2019, in which data is abundant and financial speculation is already the norm. Only a few tools are seeking to regenerate our ailing homes and buildings; most are reducing homes to data points that can be traded at the click of a button. A step change in the real estate system has arrived.<\/p>\n Martha Dillon is a writer on climate justice, housing and the built environment<\/i><\/p>\n","protected":false},"excerpt":{"rendered":" At first glance, US real estate firm Opendoor offers a pretty good deal. People looking to sell their home can upload their address, basic details and some photos to the website. It scours the information, cross-references market data and makes an offer. If the offer is accepted, the seller is sped through the transaction process, …<\/p>\n","protected":false},"author":9844,"featured_media":734600,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_oasis_is_in_workflow":0,"_oasis_original":0,"ep_exclude_from_search":false},"categories":[745],"tags":[93810,100906],"yoast_head":"\n