Artificial intelligence (AI)—once an abstraction of Silicon Valley and Hollywood—is rapidly becoming a daily tool in everything from vacation planning to research and writing, as well as mortgage underwriting to regulatory compliance and property maintenance. For the affordable housing sector, the question isn’t whether to adopt AI, but how to do it responsibly, effectively, profitably, and equitably.
At the National Housing Conference’s recent National Advisory Council meeting, we brought together leading experts on AI in housing to explore how stakeholders are using this rapidly advancing technology. NHC’s Senior Research Director Brittany Webb, who leads NHC’s adoption and implementation of AI, moderated a discussion that explored the promise, paradoxes, and potential pitfalls of this new frontier.
AI technology is evolving faster than any prior industrial or technical innovation. ChatGPT, one of several generative AI models that captured global attention in 2023, reached two billion searches in two years—nearly a decade faster than Google. By then, AI had already demonstrated its computing power, defeating GO Master Fan Hui. This wasn’t just an entertaining example of powerful computing. It is estimated that there are 10170 possible moves in the ancient Chinese game, Go (that’s a 1 followed by 170 zeros). By comparison, Chess has 10120 possible moves, and scientists believe that there are about 1082 atoms in the observable universe (one-hundred thousand quadrillion vigintillion atoms).
What was revolutionary is that DeepMind, the AI system that defeated Fan Hui, did so by teaching itself to play the game from scratch. After 500 matches with itself, it was developing strategies never played in the 2,500 year history of the game. That ability to learn and adopt is what makes AI so powerful, exciting, and frightening, as Bill Gates has been particularly outspoken about.
Webb noted that development and adoption costs—computing, storage, and model training—are falling much faster than historical declines in electricity and memory prices. Simultaneously, computing power and application is growing exponentially faster than any preceding technology.
That acceleration is already reshaping housing. Panelists discussed a range of real-world uses: AI that translates client communications in real time, reviews complex local and national regulations codes for compliance in multiple municipalities, and automates preparation and organization of compliance documents that once consumed staff time. Others are experimenting with predictive maintenance tools that flag potential system failures before tenants are impacted. These are not futuristic concepts; they’re being piloted in housing agencies, nonprofits, and private firms today.
For smaller nonprofits, AI offers an opportunity to compete with larger players. Marisa Calderon, President and CEO of Prosperity Now, emphasized that properly implemented tools can help nonprofits scale outreach and data analysis without equivalent increases in staff or budget. Her organization uses AI to analyze thousands of policy metrics in a fraction of the time, freeing human experts to focus on strategy rather than spreadsheets.
Chris Spanos, who transitioned from automotive AI innovation to housing, warned against what he called “party-trick AI”—tools that impress in demos but fail to address operational challenges at scale. He has built an AI system that tracks regulatory compliance, turning what was once a headache of spreadsheets and deadlines into an automated process. The difference, he said, lies in focusing on “practical AI” that solves specific problems, not just captures attention.
Even as automation expands, the human element remains central. Sarah Bainton Kahn described how JPMorganChase uses AI to assist, not replace, customer service staff. Large language models can help agents identify the core of a tenant’s concern and suggest responses in real time, improving both efficiency and empathy. “You want humans dealing with exceptions, and robots dealing with everything else,” as Spanos put it.
This balance is crucial for public trust. Consumers are more comfortable knowing that AI supports their experience rather than mediating it entirely. As Debra Still, Vice Chair of Pulte Financial Services and Co-Chair of the NHC National Advisory Council noted, reducing costs through automation should expand and improve human interaction where it matters most.
Jamie Dimon, CEO of JPMorganChase, has said his company recovered the cost of adopting AI — $2 billion, by the time it was implemented. The meeting also addressed AI’s physical footprint. Amazon’s new AI data center near Pennsylvania’s Susquehanna nuclear plant will draw up to 960 megawatts of power—enough to power between 500,000 and 800,000 homes. AI may be digital, but it runs on energy and water, and its environmental impact is substantial.
NHC’s adoption of AI should cost much less than JPMorganChase, but we are confident that it will yield results that far exceed our investment. We have hired Anant Corporation to help us integrate ChatGPT corporate platform into our work. Anant’s CEO, Rahul Singh, encouraged NAC members unfamiliar to try out AI for themselves on platforms like ChatGPT 5.0 or Perplexity.
Calderon warned that this imbalance could deepen the digital divide: “Large swaths of America don’t have the bandwidth—literally or figuratively—to use this technology. That builds an even bigger chasm of inequity.” Access to digital infrastructure is quickly becoming as essential to housing opportunity as access to financing.
Panelists agreed that AI’s trajectory in housing depends on governance and guidance as much as algorithms. Bainton Kahn called for a unified federal framework for AI regulation, arguing that today’s fragmented standards leave too much room for confusion and misuse.
Still, the mood was cautiously optimistic. Tim Rood, CEO of Impact Capitol, noted that AI “is only going to get better.” Used thoughtfully, AI can be a deflationary force—lowering costs, improving operations, and widening access to opportunity. He also cautioned that AI companies must be constantly innovating and adopting to the rapidly evolving technology. Impact Capitol’s Housing AI platform ALFReD is integrated into NHC’s Housing Resource Center.
The role of artificial intelligence in housing is no longer hypothetical—it’s already here. From compliance tracking to customer engagement, AI’s potential to reduce costs and close information gaps is immense. Yet, its benefits will only be realized if housing leaders ensure that efficiency serves equity. Lisa Rice, CEO of the National Fair Housing Alliance, emphasized this point during a recent Beyond Four Wall’s podcast with me. NFHA’s Tech Equity Initiative is a multi-faceted effort designed to eliminate bias in algorithmic-based systems used in housing and financial services, increase transparency and explainability for AI tools, outline ethical standards for responsible tech, advance effective policies for regulating AI tools, and increase diversity and inclusion in the tech field.
As AI reshapes industries, the housing sector has a unique responsibility to guide its adoption toward fairness, transparency, and inclusion. The technology may be new, but the mission remains the same: to ensure everyone has access to a safe, affordable home—and now, a fair share in the digital future that builds it.
____________
For a deep dive on the growth of AI, see Mary Meeker’s presentation on Trends – Artificial Intelligence from May, 2025