Machine learning or “AI” is a novel approach to solving problems with software. Computer programs are designed to take large amounts of data as inputs, and recognize patterns that are difficult for humans to see. These patterns are so complex, it is sometimes unclear what parts of the data the machine is relying on to make a decision. This type of black box often produces useful predictions, but it is unclear whether the predictions are made based on forbidden inferences. It is possible that machine learning algorithms used in banking, real estate, and employment rely upon impermissible data related to protected classes, and therefore violate fair credit, housing, and labor laws.
Machine learning, like all programmed software, follows certain rules. The software generally compares images or numbers with
descriptions. For instance, facial recognition software may take as inputs pictures of faces, labeled with names, and then identify patterns in the face to associate with the name. Once the software is trained, it can be fed a picture of a face without a name, and the software will use the patterns it has developed to determine the name of the person that face belongs to. If the provided data is accurately prepared and labeled, the software usually works. If there are errors in the data, the software will learn the wrong things–and be unreliable. This principle is known as “garbage in, garbage out”–machine learning is only ever as reliable as the data it works with.
This causes problems when machine learning is applied to data that has racial bias baked-in to its core. Machine learning has been used in the criminal justice system to purportedly predict–based on a lengthy questionnaire–whether a perpetrator is likely to reoffend. The software was, after controlling for arrest type and history, 77% more likely to wrongfully predict that black defendants would re-offend than white defendants. Algorithms used by mortgage lenders charge higher interest rates to black and latino borrowers. A tenant screening company was found to rely partly on “arrest records, disability, race, and national origin” in its algorithms. When biases are already present in society, machine learning spots these patterns, uses them, and reinforces them.
There are two ways of interpreting the racial effect of machine learning under discrimination law: disparate treatment and disparate impact. Disparate treatment is fairly cut-and-dry, it prohibits treating members of a protected class differently from members of an unprotected class. For instance, giving a loan to a white person, but denying the same loan to a similarly situated black person, is disparate treatment. Disparate impact is using a neutral factor for making a decision that affects a protected class more than an unprotected class. Giving a loan to a person living in a majority-white zip code, while refusing the same loan to a similarly situated person living in a black-majority zip code, creates a disparate impact.
Whether machine learning creates disparate treatment, disparate impact, or neither, depends on the lens the system is viewed through. If the software makes a decision based on a pattern that is the machine equivalent to protected class status, then it is committing disparate treatment discrimination. If a lender uses the software as its factor in deciding whether to grant a loan, and the software is more likely to approve loans to unprotected class members than it is to approve loans to similarly situated protected class members, then the lender is committing disparate impact discrimination. And if a lender uses software that relies on permitted factors, such as education level, yet nevertheless winds up approving more loans for unprotected class member than it does for protected class members, the process could be discriminatory but not actionable under either disparate treatment or disparate impact theory.
The problem thus depends in part on understanding how the machine interprets the data–the exact information we often lack when machine learning is particularly complex. Thus, instead of relying on either disparate impact or disparate treatment theory, perhaps legal analysis of discrimination in machine learning should be entirely outcomes-driven. If, in fact, an algorithm wrongly predicts the likelihood of an event occurring, and that algorithm is less accurate for protected class members than unprotected class members, the algorithm should be considered prima facie discriminatory. Such a solution is viable for examining recidivism, interest rates and loan repayment, but it may be insufficient to cover problems like housing denials or employment. If an algorithm denies protected class members access to housing in the first place, it is difficult to falsify the algorithm’s decision, as nonpayment or other issues necessarily do not arise.
Machine learning offers great promise to escape the racial biases of our past if it is programmed to only rely on neutral factors. Unfortunately, it is currently impossible to determine which factors used by machines are truly neutral. Improperly configured machine learning has led to higher incarceration rates for black inmates and higher loan interest rates for black and latino homeowners. Such abuses of machine learning technology must be rejected by the law.
COVID-19’s impact on the world has been unsparing: it has taken thousands of loved ones and overburdened healthcare systems. Its economic impact has been no less devastating. Government-mandated shutdowns have forced many small businesses to shut their doors. Those lucky enough to remain open have seen a marked decrease in customers. A McKinsey & Company survey polling small and medium-sized businesses found that 59% of small or medium-sized businesses reported experiencing a “significant” drop in personal and business income. 25% had faced such a precipitous drop that they expected they would have to file for bankruptcy in the coming months. This impact will have a broad effect: these businesses represent 48% of the U.S. economy and employ over 60 million people.
COVID-19 has also been unsparing when it comes to consumers. Not only are those employed by small or mid-sized businesses losing their jobs, but even those working at large international companies are also losing their jobs or facing pay cuts. Self-employed individuals have seen their business opportunities vanish. Right now, because of moratoria on foreclosures, evictions, and debt collection, these issues are looming when the overall economy begins to resume. Without any intervening forces, a wave of evictions, foreclosures, repossessions, and bankruptcies is likely to come.
Currently, little protects consumers and small businesses during emergencies or disasters. While larger institutions are expected to have plans, even participating in stress tests with regulators, small businesses usually do not have such protective measures. Indeed, because so many operate on a thin profit margin, investing in plans to address something that seems unlikely to ever happen may seem like a waste of money. Moreover, consumers are not always going to be financially prepared either. The average American has $8,863 in their savings account. This number decreases with age: for example, young single adults under 34 have an average of $2,729 in savings. There are racial disparities too, with an average of $1,500 for Hispanic and $1,000 for Black savings account balances. (Note that this data does not include the unbanked, which is largely composed of Black and Hispanic households.)
In March 2020, Senator Sherrod Brown (D-Ohio) introduced the Small Business and Consumer Debt Collection Emergency Relief Act of 2020 (the “Act”). The Act amends the Fair Debt Collection Practices Act (“FDCPA”) to provide temporary forbearance periods after “major disasters or emergencies.” The Act could provide much-needed relief to small businesses and American consumers by expanding on current protections currently provided by the FDCPA.
The FDCPA became law in 1978. Its purpose was to crack down on abusive debt collection tactics. In drafting the law, Congress found that abusive debt collection practices led to “personal bankruptcies, to marital instability, to the loss of jobs, and to invasions of individual privacy.” The FDCPA covers several types of consumer debts: mortgage, credit card, student loan, and medical debts, just to name a few. Debts owed by businesses are currently not covered by the FDCPA.
One of the biggest protections afforded under the FDCPA is limiting how and when a debt collector can contact a debtor. Debt collectors cannot harass debtors, misrepresent what can occur if the debtor fails to pay, or contact the debtor in the middle of the night. While the FDCPA is imperfect—consumers regularly report violations—it provides an avenue of recourse and is regularly enforced.
The Act’s Proposals
The Act seeks to amend the FDCPA in several ways. First, the Act amends section 3 of the FDCPA to restrict debt collection during a national disaster or emergency. The Act defines “national disaster or emergency” two ways: (1) one declared by the president under section 401 of the Robert T. Stafford Disaster Relief and Emergency Assistance Act; or (2) “an emergency involving Federal primary responsibility that is determined to exist by the President” under section 501 of that act. Next, the Act prohibits debt collectors from numerous actions ranging from charging higher fees or interest rates to seizing the property or asset at issue. Moreover, the Act pauses any legal action against the debtor, even if the proceeding began before the disaster was declared. Finally and crucially, the Act extends FDCPA protections during national disasters or emergencies to small businesses.
The Act is a way of addressing some of the consumer protection shortcomings of the CARES Act. For example, there is a 120 day moratorium on foreclosures and evictions, but only for those who have federally-backed mortgages, live in buildings financed by federally-backed mortgages, or receive rental funding through a federal program. While this covers a significant number of individuals and homes, about 30% of Americans are left out of this provision. The debt collection moratorium under the Act would extend to those Americans. It also addresses the issue the COVID-19 pandemic has made clear: small businesses suffer when they are forced to shut down, as they may not always have the resources or option to continue operating virtually. Their financial troubles largely came as the result of a forced shutdown. More generally, the Act’s amendments can cover future national emergencies or pandemics.
While this is not an overall solution to COVID-19’s economic impact, it can provide some relief to consumers and small businesses. It may allow people to be more understanding of opening up gradually rather than quickly, too. While it does not solve a reduced or lack of income, having some assurance that the car will not be repossessed or a debt collector will not be regularly calling can provide some much needed relief in what has been a stressful time. Currently, debtors do not have such relief. Despite the pandemic, debt collectors have still been calling consumers. In Massachusetts, the Attorney General’s Office sought to ban debt collection during the height of the pandemic. However, a federal judge ruled against the state’s emergency resolution. FTC guidance is not particularly helpful either: it recommends talking things out with the debt collector and asking that they stop calling. Congress has already done a lot through the CARES Act to help consumers during the pandemic. To further assist them, Congress should pass Senator Brown’s bill.
There is no question that Massachusetts generally, and the Boston area in particular, is mired in an affordable housing crisis. Renters and would-be buyers are heavily burdened with disproportionate housing costs and are forced to compete for a limited supply of units. As of 2019, Massachusetts rental costs climbed to the third highest in the nation, surpassed only by California and Hawaii, making this the rare occasion in which beating New York isn’t something to celebrate. To increase housing, the state must address the burdensome rules that hinder the development of new units; especially multifamily homes. One possible solution is the proposed Housing Choices Act.
Though there are many facets to this issue and no simple solution, Gov. Charlie Baker is focusing on attacking the restrictive zoning rules and arduous permitting processes. Gov. Baker argues these rules are causing shortfalls in housing stock, driving up prices as demand far outstrips supply. Gov. Baker’s bill, the Housing Choices Act, was stalled throughout 2019, until late December when it was referred to the House committee of Ways and Means.
The Act has garnered the support of organizations such as the Boston Medical Center, Real Estate Bar Association, Massachusetts Municipal Association, Massachusetts Association of Community Development Corporations, Metro Housing Boston, and many more.
Gov. Baker’s administration has set a target of building 135,000 new housing units by 2025 and the Governor believes rezoning to allow more multifamily construction is a key piece in increasing housing production. Since 2017, Gov. Baker’s Housing Choices Initiative has incentivized municipalities to help meet this target by providing grants to support housing construction, but access to those grants still requires communities to vote in favor of rezoning. The Housing Choices Act would simplify rezoning by reducing the threshold of votes needed to pass certain kinds of progressive, production-oriented zoning changes in Massachusetts municipalities from two-thirds to a straight majority vote. Massachusetts is among the few states which require a supermajority for these kinds of zoning votes. The Governor calls the Housing Choice Act a “critical” piece of legislation in pursuit of production goals because it will ease that rezoning process.
Among the zoning changes to which the new straight majority vote would apply are the adoption of “smart-growth” districts under Ch. 40R, which governs…. “Smart growth” includes the construction of multi-family buildings, special permits for high-density construction, the reduction of size and parking requirements, transit-oriented projects with a set percentage of affordable units, and permission to construct accessory dwelling units.
Criticism: Too Narrow or Too Broad?
Critics such as Darnell Johnson, regional coordinator at Right to the City Boston, however, believe that the bill “ignore[es] the needs of working families” and will only lead to more luxury condos and other housing out of reach for most Massachusetts residents. Chris Norris, executive director Metro Housing Boston, concurs, and points out that with 250,000 low-income households forced to spend upwards of 50% of their income on rent, “incentives alone are not sufficient to product affordable housing.” Norris cites the fact that Ch. 40R was passed in 2004, was expected to generate 33,000 new units over ten years, and more than $20 million has been poured into its incentives. As of 2017, however, this law has produced only 3,500 units, less than half of which are designated as affordable.
Advocates within and without the Baker administration counter that this bill is not meant to be a panacea for housing woes but rather a step in the right direction. Clark Zeigler, executive director of the Massachusetts Housing Partnership, points out that right now good housing proposals are defeated despite winning majority support because of the current two-thirds voting requirements. Such was the case in Salem, where a measure to increase accessory dwelling units, one of the Housing Choice Act’s targets, won the majority, but not supermajority, and thus failed to pass. Salem Mayor Kim Driscoll supports Gov. Baker’s efforts to prevent situations like this, calling on local authorities to pursue a “strong partnership with state leaders.” Once this problem is solved, MHP and other affordable housing advocates can and will continue pushing for further means of providing relief to those overly burdened by housing costs. In an official statement, CHAPA (Citizens’ Housing and Planning Association) made its support of the Housing Choice Act clear for the same reasons. “Lowering the super majority threshold required for zoning changes will empower the simple majority of people in a community to vote ‘Yes’ for housing.” CHAPA sees the bill as an essential opportunity to help communities “encourage housing development and undo policies that prevent housing production and perpetuate segregation.”
Another source of criticism is a fear that the bill would inevitably cause unwanted and overly dense construction which will damage the character of older neighborhoods. Similarly, Rep. Smitty Pignatelli (D-4th Berk.) proposes that the switch to straight majority be predicated on local approval out of concern for state-level infringement on local autonomy. Sen. Brendan Crighton (D-Lynn), co-chairman of the Housing Committee, does not believe this is a real cause for concern, insisting that the bill “doesn’t take away any local control. . . [but rather] puts together a more common-sense approach that says a majority of folks can make a decision on the matter.” He sees this and similar legislation as crucial because Massachusetts’ housing production is half of what it was in the 1970s while rent has risen by 75% over the past twenty years.
Some communities, such as Needham and Springfield, believe they and some 70 other municipalities should be exempt from the bill because 10 percent or more of their housing stock already meets state “affordability” thresholds. Needham Select Board Chairman John Bulian sees requiring his town to comply would be penalizing it despite its having exceeded Massachusetts’ affordable housing percentage goals. Such complaints, however, seem to miss the larger picture of spurring more housing growth by enabling local residents greater control over their zoning through simplified zoning vote processes.
Massachusetts is in a housing crisis—one that can only be fixed by finding a way to build more housing. An important first step is to reform the rules and processes that restrict the needed housing units.
Massachusetts’s Governor Baker signed An Act Regulating and Insuring Short-Term Rentals on December 28, 2018. The act regulates short-term rentals provided through services like Airbnb. The governor rejected an earlier version of the bill, and returned amendments primarily allowing for an exemption for owners who rent out their property for two weeks or less per year, and reducing the amount of information provided publicly about rentals owners. The bill was motivated by concerns that the rise in short-term rentals drives up housing costs and pushing out long-term residents. The statewide bill comes after both Boston and Cambridge individually passed laws essentially having the same effects. However, the Boston law was challenged by Airbnb, who filed suit in federal court claiming that the regulations are “Orwellian” and violate several laws, including laws that protect online companies from being held liable for the actions of their users. The city of Boston is currently holding off on some of the regulations passed pending the resolution of the court case. Airbnb had not yet said if it will challenge the new Massachusetts law.
The statewide law has two main components: first, that all short-term rentals are taxed by the Commonwealth, and can be additionally taxed by local governments, and second, that all owners of short-term rental properties must register with the state and hold insurance. The registration requirement was a cause of debate. Lawmakers, including Governor Baker, were concerned about violating the privacy of owners by publishing their names and addresses publically. In the amendments to the July bill that Governor Baker rejected, the registration requirement was changed so that only the owner’s neighborhood and street name would be published, not their exact address. The law also dictates eligibility in order to register. To be eligible to be a short-term housing unit, the space must be compliant with housing code, be owner-occupied and be classified for residential use, among other requirements. Another cause of debate was an exemption for occasional renters. Governor Baker originally wanted owners who rent their properties for 150 days or less to be exempt from the regulations. However, in the final version of the bill, the exemption was decreased to 14 days.
Not surprisingly, the hotel industry supports the bill. Paul Sacco, the President and CEO of Massachusetts Lodging Association, said:
“This is a tremendous victory for municipal leaders and the people of Massachusetts who have been waiting for years while Airbnb rentals have exploded, resulting in skyrocketing housing costs and disruptions in local neighborhoods. By adopting a more level playing field between short-term rentals and traditional lodgers, lawmakers made great strides toward a more fair and sensible system.”
Airbnb had a far less enthusiastic response however. In citing concerns about the property owners who use Airbnb to earn extra income, Airbnb said that they would “continue the fight to protect our community and the economic engine of short-term rentals for hosts, guests, and local small businesses”
While Massachusetts is the first state to pass a law, many other cities have passed similar laws in the recent years. In Nashville, the city passed a law in January which focused on taxing short-term rentals that are not owner-occupied in order to fund affordable housing development in the city. The “linkage fee” tax is controversial, with lawmakers questioning if the fees generated are enough to actually impact the lack of low-income housing within the city. Seattle passed a similar tax law in November 2017. The legislation aims to encourage owners who rent out a spare bedroom and discourage investors who buy entire buildings for use as short-term rentals. Finally, New York City passed regulations in July 2018 which requires Airbnb, and other similar companies, to provide information about the properties listed for rental within the city. However, Airbnb sued in federal court claiming that the requirement to provide information to the city violated the company’s fourth amendment right against search and seizure. The law was set to take effect in February 2019, however the Judge granted a preliminary injunction in favor of Airbnb saying, “The city has not cited any decision suggesting that the government appropriation of private business records on such a scale, unsupported by individualized suspicion or any tailored justification, qualifies as a reasonable search and seizure.”