More landlords and lenders are using AI. Fewer regulators are checking them for bias.

The housing industry is experiencing a significant transformation as artificial intelligence (AI) tools become increasingly prevalent in determining who qualifies for home loans or leases. However, this surge in AI adoption coincides with a rollback of long-standing civil rights protections aimed at ensuring fairness in these evaluations. This article explores the implications of these developments and the potential risks associated with unregulated AI in housing.

The Rise of AI in Housing

AI technology is not new to the housing and lending sectors. Algorithms have long been utilized to predict various outcomes, such as a home’s selling price or the likelihood of an individual affording their rent. Recent advancements in AI have made these tools more accessible, leading to heightened interest among mortgage and real estate businesses in employing computerized systems.

Proponents of AI argue that these technologies can offer a more objective approach to housing decisions, potentially reducing discriminatory bias and addressing entrenched inequalities. However, concerns have arisen regarding the data used to train these AI models, which often reflect historical patterns of discrimination.

Concerns Over Bias in AI

Critics warn that while AI has the potential to advance civil rights, it may also reinforce existing societal discrimination if not carefully monitored. Federal Reserve Governor Michael Barr has voiced concerns, stating, “Artificial intelligence might advance civil rights if it’s used properly… but it might also reinforce discrimination in our society if we’re not careful.”

Despite assurances from technology developers that they train their systems to avoid bias, the lack of robust government oversight raises alarms. The diminishing enforcement of anti-discrimination regulations could weaken the incentives for companies to prioritize fairness in their AI systems.

The Impact of Regulatory Rollbacks

Since taking office, the Trump administration has actively sought to limit the federal government’s ability to enforce rules based on “disparate impact.” This legal standard evaluates whether a practice results in discrimination against a group, regardless of the intent behind it. Historically, disparate impact has been a crucial tool for challenging decisions influenced by algorithmic technology.

In 2024, a federal court awarded over $2 million to rental applicants who alleged they were denied housing due to an algorithm that disproportionately affected Black and Hispanic individuals. The plaintiffs successfully argued that the algorithm overly relied on credit scores without considering other factors, such as housing vouchers that could enhance applicants’ ability to pay rent.

Shifts in HUD’s Stance

Under the first Trump administration, the Department of Housing and Urban Development (HUD) acknowledged the importance of disparate impact methods in identifying potential discrimination in algorithmic systems. However, in the second Trump term, HUD’s position shifted, suggesting that enforcing disparate impact standards was unfair to businesses and led to illegal racial preferences.

HUD spokesperson Robbie Myers stated, “The issue isn’t AI – it’s disparate impact, a discredited theory that requires individuals and entities to consider race on the front end to avoid legal liability on the back end.” This change in narrative has raised concerns among civil rights advocates who argue that such enforcement is essential for challenging harmful housing provider decisions influenced by algorithms.

The Role of Individual Lawsuits

While disparate impact methods remain applicable under existing civil rights laws, individuals can still use them to file lawsuits. However, the lack of transparency in many AI tools can make these cases challenging for non-experts to navigate. Lisa Rice, president of the National Fair Housing Alliance, emphasized that bringing complaints can be labor-intensive and costly for typical consumers.

Rice argues that relying solely on individuals to enforce disparate impact standards is insufficient. She advocates for government agencies with the necessary expertise and resources to investigate these systems and compel institutions to rectify discriminatory algorithms.

Industry Perspectives

Despite concerns about discrimination, many industry advocates support the rollback of disparate impact rules. Organizations like the Community Home Lenders of America have expressed that previous regulations were overly expansive and misaligned with past court decisions. They argue that excessive oversight could hinder the adoption of AI tools that benefit small businesses and reduce human bias.

Rob Zimmer, a spokesperson for the association, stated, “Let’s make sure we have a system that evaluates people based on math — not based on personal characteristics that arguably should have nothing to do with whether or not you recommend access to credit.” He emphasized that the policy changes were necessary to prevent the stifling of innovation in the industry.

The Future of AI in Housing

As the Biden administration focuses on preventing discriminatory outcomes in housing technologies, some experts warn that an overemphasis on regulation could inadvertently harm the communities it aims to protect. The balance between fostering innovation and ensuring fairness in housing decisions remains a contentious issue.

In conclusion, the rapid adoption of AI in the housing industry presents both opportunities and challenges. While AI has the potential to enhance objectivity in lending and leasing decisions, the rollback of regulatory protections raises significant concerns about the potential for discrimination. It is crucial for stakeholders to engage in a dialogue about the ethical implications of AI in housing and the importance of maintaining robust oversight to protect vulnerable communities.

Note: The information presented in this article reflects the current state of AI adoption in housing and the regulatory landscape as of October 2023. Ongoing developments may further influence these dynamics.