Maryland Tenant Screening Laws: What Landlords Using AI Must Change

1. What changed in Maryland tenant screening laws on October 1, 2026?

Maryland tenant screening laws changed on October 1, 2026, when three 2026 statutes took effect together: a disparate impact standard for covered housing practices, a fair chance process for criminal history checks, and new limits on how credit and income are used to screen applicants who pay with a housing subsidy. All three were approved by Governor Wes Moore on May 26, 2026.

The headline change is HB 573, enacted as Chapter 778 of 2026. It amends Section 20-705 of the State Government Article so that the existing housing discrimination prohibitions apply “regardless of intent,” and it adds a new sixth prohibition: a person may not “act in a manner that has a discriminatory effect.” The bill passed the House 98 to 34 and the Senate 33 to 12 on final passage on April 13, 2026, according to the votes posted on the General Assembly’s bill page.

Maryland’s new tenant screening laws are three October 1, 2026 statutes that change how landlords may screen rental applicants. HB 573 bans covered housing practices with an unjustified discriminatory effect, the Maryland Fair Chance Housing Act moves most criminal history review behind a conditional offer, and HB 315 bars credit score rejection of subsidized applicants and limits income tests to their own rent share.

None of these laws uses the words artificial intelligence, algorithm or automated decision. That is the point worth understanding. They are written around outcomes and process steps, so a rule embedded in tenant screening software is judged the same way as a rule written in a leasing manual. If your screening runs through a vendor’s score, a property management platform’s default settings, or an AI assisted application review, these are among the Maryland housing rules that apply. Privacy, consumer reporting, consumer protection and local requirements can apply as well.

Readers who follow our coverage will recognize the pattern. Colorado took a disclosure approach to automated housing decisions, which we covered in our guide to the Colorado ADMT Act for tenant screening. Illinois codified disparate impact for lending, which we covered in our article on Illinois AI underwriting. Maryland did the housing version, and it is already in force.

2. What does Maryland’s disparate impact law, HB 573, prohibit?

Maryland’s disparate impact law prohibits a covered housing practice that “actually or predictably” has a disparate impact on a protected group, or that creates, increases, reinforces or perpetuates segregated housing patterns, unless the practice meets a three part justification. The definition sits at Section 20-701(b-1) of the State Government Article, and the prohibition in Section 20-705 remains subject to the existing exemptions in Sections 20-703 and 20-704.

HB 573 expressly codifies discriminatory effects liability; it does not mark the first recognition of disparate impact claims under Maryland housing law. The Supreme Court of Maryland had already allowed such a claim to proceed in July 2025, as discussed in section 3.

The definition covers 12 protected characteristics: race, color, religion, sex, disability, marital status, familial status, sexual orientation, gender identity, national origin, source of income and military status. That is five more than the seven classes in the federal Fair Housing Act. Source of income matters most for screening, because Maryland has protected it since the 2020 HOME Act, and income and credit rules are where screening tools do most of their sorting.

The justification is in new Section 20-705(b). A person has not violated the section if all three of the following are true:

• the action was without discriminatory intent;
• the action was necessary to achieve one or more substantial, legitimate and nondiscriminatory interests; and
• the person could not have achieved those interests by less discriminatory means.

Two drafting points matter for landlords. First, the defense is conjunctive. A legitimate reason is not enough on its own if a less discriminatory method would have served it, which is the question that decides most screening disputes. Second, the statute does not say which side carries the burden of proof on each element. Courts will fill that gap, and the Supreme Court of Maryland has already described the federal three step framework in this context, as discussed in section 3.

HB 573 also authorizes the Department of Housing and Community Development to adopt regulations on affirmatively furthering fair housing, but that authority is aimed at the department, counties and housing authorities, not private landlords. The Department of Legislative Services fiscal and policy note says the bill “codifies and affirms the disparate impact analysis currently available in housing discrimination cases” and reports that the Maryland Commission on Civil Rights received 266 housing discrimination cases for investigation in fiscal 2025.

3. Do Maryland tenant screening laws apply to AI screening scores?

Yes. Maryland tenant screening laws apply to AI screening scores and automated screening rules because the disparate impact test looks at what a practice does, not at who or what performed it. A landlord’s selection or use of a score cutoff, an automated recommendation or a platform default may be a challenged screening practice. Responsibility depends on how the system was selected, configured and used, but outsourcing does not by itself eliminate exposure.

The clearest Maryland illustration predates HB 573. In Hare v. David S. Brown Enterprises, Ltd., 491 Md. 653, filed July 28, 2025, a voucher holder applied for an apartment renting for $1,590 a month. Her voucher would have covered $1,464, leaving her responsible for $126. The landlord had her submit income information to a third party that evaluated whether she met a minimum income requirement of 2.5 times the full rent. The third party combined her $841 in monthly Supplemental Security Income with the voucher amount, found that the $2,305 total fell short of the required $3,975, and the landlord rejected her. The opinion does not describe the third party process as AI or machine learning; the case is a lesson about outsourced policy design, not evidence of algorithmic discrimination.

The court agreed the landlord had not engaged in disparate treatment, because it counted voucher income the same way as every other source. But it vacated the judgment and sent the case back, holding that the disparate impact claim still had to be decided. The question, the court wrote, is whether a facially neutral requirement “falls more harshly on voucher holders” than an appropriately comparable group and, if so, whether the landlord can identify a legitimate business need. The opinion did not find the landlord liable. It held that applying a rule evenly is not a defense to an effects claim.

The federal example is Louis v. SafeRent Solutions, No. 1:22-cv-10800 in the District of Massachusetts. Two Black applicants holding vouchers alleged that SafeRent’s screening score disproportionately penalized Black and Hispanic renters. The court granted final approval of a $2.275 million settlement on November 20, 2024. According to the Civil Rights Litigation Clearinghouse summary of the court file, SafeRent agreed for at least five years to stop issuing accept or decline recommendations for voucher holders based on its scoring model unless the model is validated for fairness, and the property manager that used the score was a party to the settlement.

How common is this exposure? TechEquity Collaborative’s April 2025 research, which surveyed landlords and roughly 2,200 tenants in California, Georgia and North Carolina, reports that 60 percent of surveyed landlords received AI enabled tenant screening reports and that only 3 percent of surveyed tenants could name the screening company that assessed them. Those are the study’s own figures, not Maryland data, but they describe a market where the decision logic is often invisible to both sides of the application.

One limit is worth stating plainly. A statistical gap alone does not establish a violation. Under the framework Hare described from the U.S. Supreme Court’s 2015 Inclusive Communities decision, a plaintiff must show a robust causal connection between a specific policy and the disparity. HB 573 supplies an express state statutory definition and prohibition, building on the disparate impact analysis Maryland courts had already recognized.

4. How does Maryland’s standard compare with the federal rule HUD wants to remove?

Maryland’s standard closely tracks the federal discriminatory effects rule at 24 CFR 100.500, but it covers more protected classes, it is a statute rather than an agency regulation, and it is not affected by HUD’s pending proposal to delete the federal rule.

HUD published a proposed rule on January 14, 2026 (91 FR 1475) that would remove Section 100.500 and leave interpretation of disparate impact liability under the Fair Housing Act to the courts. On August 10, 2026 it published a supplemental proposal (91 FR 51416) that would also remove disparate impact provisions from its Title VI regulations, with comments due October 9, 2026. The eCFR version we reviewed, current through October 30, 2026, still contained Section 100.500, and our Federal Register search on November 3, 2026 found no final removal rule.

Even a final federal rule would not remove disparate impact claims under the Fair Housing Act itself. The Supreme Court recognized those claims in 2015, and HUD’s own proposal says it would leave the question to the courts. A final rule would remove HUD’s regulatory formulation and could affect agency enforcement and litigation analysis, but it would not overrule the Supreme Court. Maryland’s statute is unaffected either way.

The source of income row is the one most likely to touch screening software. A tool that scores applicants partly on income stability, income type or credit file depth can produce very different outcomes for voucher holders, and voucher status is a protected characteristic in Maryland even though it is not one under federal law.

5. What does the Maryland Fair Chance Housing Act require?

The Maryland Fair Chance Housing Act requires covered landlords to delay most criminal history review until after a conditional offer, give written notice before taking an application fee, and conduct an individualized assessment before withdrawing an offer. It is SB 937, enacted as Chapter 752 of 2026, and it adds a new Subtitle 2A to Title 8 of the Real Property Article.

Coverage turns on portfolio size. The subtitle applies only to a landlord that manages or owns five or more residential rental units in Maryland, counting units owned or controlled in whole or in part, directly or indirectly, or through one or more legal entities. It does not apply to owner occupied rental units. Splitting units across LLCs does not avoid the count. Note that some summaries, including a state web page, describe the threshold as “more than 5”; the enacted text says “five or more.”

The required workflow has five parts:

• Before the application fee. Give written notice that certain convictions may be considered only after a conditional offer, explain the applicant’s right to provide evidence of inaccuracies, rehabilitation and mitigating factors, and say an individualized assessment will be conducted.
• Before the conditional offer. No criminal history question on the application and no oral or written inquiry, except for the specified serious conviction and sex offender registration exceptions listed in Section 8-2A-05. Those include, among others, first and second degree murder and human trafficking. Configure them against the statutory list, not a vendor’s generic “violent crime” or “felony” category.
• After the conditional offer. The landlord may run a check and consider felony convictions within the 5 years before the rental application, equivalent out of state or federal felonies in the same window, and any listed conviction the applicant did not disclose.
• Withdrawal. An offer may be withdrawn only if the landlord determines withdrawal is necessary to fulfill a substantial, legitimate and nondiscriminatory interest. The written notice must give a specific reason and explain how to request reassessment with written evidence signed under penalty of perjury.
• Records. Within 30 days of a withdrawal notice the applicant may request everything the landlord relied on, and the landlord must provide it within 10 days.

On reassessment the landlord must weigh six factors, including the nature and severity of the offense, the applicant’s age at the time, time elapsed, rehabilitation evidence, safety risk if it recurred, and whether it occurred on the rented property. A landlord that conducts criminal history checks under the Act must do so for every prospective tenant, not selectively. The Act also bars requiring applicants to take a drug or alcohol test, bars requesting or requiring consent to release information from a substance use disorder prevention or treatment program, bars advertisements that say people with records will not be considered, and preempts comparable local ordinances. The Department of Housing and Community Development published three optional model forms on October 6, 2026: a pre application notice, an individualized assessment worksheet and a denial letter.

The automation problem is straightforward. Many screening packages run credit, eviction and criminal checks in a single pull at submission and return one combined recommendation. For a covered Maryland landlord, an unrestricted pre offer criminal history pull is incompatible with the general sequencing rule. The system has to distinguish the limited permitted pre offer inquiries from the broader post offer check, narrow the post offer review to the permitted categories and lookback, and route any withdrawal through the individualized assessment rather than an automated decline.

6. How does HB 315 change credit scores and income ratios for voucher holders?

HB 315 restricts income based rejection of subsidized applicants, bars rejection based on credit score or lack of a credit score, and protects adverse credit history from a confirmed period before the subsidy, while still permitting an income test on the applicant’s own rent share. It applies to a Maryland landlord that uses financial information, including credit history, in screening an applicant who pays rent with an income based housing subsidy. It is Chapter 772 of 2026, adding Section 20-705.1 to the State Government Article, and a violation is a discriminatory housing practice enforced by the Maryland Commission on Civil Rights.

An income based housing subsidy is recurring assistance paid to the landlord by a government entity or nonprofit organization to defray rent, including Housing Choice Vouchers, so the protection is broader than Section 8 alone. The pre subsidy credit protection applies when the period without the subsidy is confirmed by the applicant, voucher paperwork, public housing authority documents, or a housing authority’s response to the landlord’s request, and the landlord may not insist on housing authority verification of the dates. The section does not expressly protect adverse credit arising while the applicant had the subsidy, although other fair housing rules still apply.

Unlike the Fair Chance Act, this provision has no unit threshold, although the general Section 20-704 exemptions still need to be assessed. The exceptions are specific. A landlord may still verify that the tenant has enough income to pay the portion of rent not covered by the subsidy, using a ratio substantially equivalent to the one used for unsubsidized tenants. A landlord may also refuse based on a commercially reasonable and nondiscriminatory use of a reference from a previous or current landlord, or a history of lease violations, unpaid utilities, nuisance or property damage. Landlords whose public or nonprofit funding requires income qualification may collect financial information, and anything authorized under federal law is preserved.

Applied to the Hare facts, the difference is stark. A 2.5 times ratio applied to the tenant’s $126 share requires $315 a month in income. Ms. Hare had $841. Applied to the full $1,590 rent, it required $3,975. The statute now settles which method Maryland expects for subsidized applicants.

The same chapter adds a positive rent reporting duty at Real Property Section 8-208.4 for landlords that own six or more residential rental units in Maryland. The written lease must include the option to have complete and timely rent payments reported to at least one consumer reporting agency. For leases entered into on or after October 1, 2026 the offer is made at signing and at least annually after that; existing tenants must receive it by January 1, 2027, by first class mail with a certificate of mailing, tracked delivery, or electronic delivery if the tenant has elected it. The offer must follow the form set by regulation and include the agency names, any fee, opt in and opt out instructions and a signature block, and an offer sent by first class mail must include a self addressed, stamped return envelope. Any fee is capped at the lesser of actual cost or $10 a month, is not rent, and its payment or nonpayment may not itself be reported. The Office of Tenant and Landlord Affairs has posted guidance on both requirements.

7. Which screening settings carry the most risk?

The screening settings that carry the most risk under Maryland’s new laws are automatic declines, bundled background checks, credit score cutoffs applied to voucher holders, and income ratios calculated on full rent. The table below is our reading of the statutes against common tenant screening software features; it is not agency guidance.

Human review does not cure a discriminatory input. If the score itself drives a disparity, a reviewer who almost always accepts it has not changed the effect. What review does provide is a place to apply the less discriminatory alternative, such as counting only the tenant’s share of rent, and a record showing you considered one.

8. Who enforces these laws and what are the penalties?

The Maryland Commission on Civil Rights enforces HB 573 and HB 315, private plaintiffs can sue in state court, and the Attorney General has independent authority to investigate civil rights violations. The Fair Chance Act carries its own civil penalty of up to $500 per violation.

For the fair housing provisions, an aggrieved person may file a complaint with the Commission, and private state court actions are subject to the 130 day waiting rule, the 2 year limitations period running from the later of the occurrence or termination of the practice, tolling while administrative proceedings are pending, and the other procedural limits in Section 20-1035, including the bar on a separate suit after a consented conciliation agreement or once an administrative hearing has begun. A court that finds a discriminatory housing practice may award actual and punitive damages, grant injunctions, and award reasonable attorney’s fees and costs. Where the Commission finds a pattern or practice raising an issue of general public importance, it may bring its own action under Section 20-1036, in which a court may assess civil penalties of up to $50,000 for a first violation and up to $100,000 for a subsequent violation. Under Section 20-1041 the Attorney General may investigate, prosecute and remediate civil rights violations on behalf of Maryland residents. That is independent public enforcement: the section bars the Attorney General from bringing such an action on behalf of an individual and excludes actions against governmental units.

The Fair Chance Act’s $500 per violation civil penalty is in the enacted text, but the Act does not spell out the enforcement route in detail. An earlier version would have made violations an unfair or deceptive trade practice under the Consumer Protection Act; that amendment was struck, although the required model notice still refers to filing a complaint under Section 13-401 of the Commercial Law Article. Check complaint routing against current DHCD guidance and the implementing regulations rather than inferring it from earlier bill versions.

Small landlords should not assume an exemption. Maryland’s existing exemptions under Section 20-704 are narrow: the sale or rental of a single family dwelling without a broker or a discriminatory advertisement, and, for a limited set of characteristics including source of income from vouchers, rooms or units in an owner occupied building with not more than five rental units. The owner occupied exemption does not cover race, national origin or disability claims, and the single family exemption is lost once the owner uses a broker or a person in the business of renting dwellings. A property exempt under one of these laws is not automatically exempt under the other two or under federal law. Industry groups have read the law the same way: the National Apartment Association’s summary describes HB 573 as permitting challenges based on discriminatory effect “even without evidence of intentional discrimination.”

Federal exposure has not gone away either. Fair Housing Act private suits remain available, and the Fair Credit Reporting Act still requires an adverse action notice whenever a consumer report contributes to a denial or worse terms. A Maryland denial letter now may need to satisfy both FCRA and the Fair Chance Act’s specific reason requirement at once.

9. What should a small landlord ask its screening vendor?

A small landlord should ask its screening vendor what the score measures, whether it has been tested for disparate impact on Maryland’s protected classes, and whether its Maryland configuration supports the two stage criminal check and voucher specific income math. Get the answers in writing; they help you explain your practices if a complaint arrives.

We suggest seven questions:

1. What data inputs feed the score or recommendation, and does it use eviction filings, criminal records, income type or credit file depth?
2. Has the model been tested for disparate impact on race, national origin, disability, familial status and source of income, and can we see a summary of the results and the date?
3. Can the platform calculate income ratios on the tenant’s share of rent when a subsidy is present, suppress credit score cutoffs for subsidized applicants, and set aside adverse credit from a confirmed pre subsidy period?
4. Can criminal history be withheld until we record a conditional offer, with only the statutory pre offer exceptions screened earlier, limited to the permitted post offer categories and lookback, and excluded from the overall recommendation?
5. Can we see the reason codes behind any decline, in language specific enough for a Fair Chance withdrawal notice?
6. What happens to applicant data, including criminal records, after a decision, and who else receives it?
7. Will you notify us when the model or its inputs change, and will you cooperate in defending a complaint?

These questions map directly onto the documentation habits in the NIST AI Risk Management Framework and the third party oversight expected under ISO/IEC 42001. A small landlord does not need a certified management system. It does need an inventory of the tools that touch applicant decisions, a written reason for each setting, and evidence of the less discriminatory alternatives it considered. The same discipline applies to rent setting software, which we covered in our article on state algorithmic rent pricing bans.

10. A 60 day plan for Maryland landlords

Because these laws took effect on October 1, 2026, Maryland landlords should pause or correct clearly incompatible practices immediately and use the following 60 days to complete the inventory, vendor review, training and documentation. This plan is a starting point, not a safe harbor.

Now: stop the clear problems. Remove criminal history questions from applications and listings except the statutory pre offer exceptions, and stop any unrestricted criminal pull before a conditional offer. Turn off credit score rejection rules for applicants with an income based housing subsidy and apply any income ratio to the tenant’s own rent share. Stop any drug or alcohol testing and any request for substance use treatment records. Check what your chatbot says about eligibility.

Days 1 to 15: inventory and count. Assess the two thresholds separately. The Fair Chance Act covers a landlord that owns or manages five or more Maryland rental units, counting units held directly, indirectly or through other entities; positive rent reporting applies to a landlord that owns six or more. Do not assume the counting rules are identical. List every tool that touches an application: screening vendor, property management platform, leasing chatbot, income verification service and identity verification service, and export the current criteria and default settings from each.

Days 16 to 30: formalize the process. Configure the two stage criminal history process, apply it to every applicant if you run checks at all, and adopt the DHCD pre application notice, assessment worksheet and denial letter, or your own equivalents. Prepare positive rent reporting language for new leases and the offer to existing tenants.

Days 31 to 45: test and justify. Send the seven vendor questions. For each remaining criterion, write one or two sentences on the interest it serves and the less discriminatory alternatives considered. Where you have enough application volume, compare approval rates for voucher and non voucher applicants as a basic check.

Days 46 to 60: lock it in. Train leasing staff on conditional offers and reassessment requests. Set calendar reminders for the January 1, 2027 rent reporting offer to existing tenants and for an annual criteria review. Keep the file. Dated records can support your explanation if a complaint arrives, but documentation does not cure a prohibited practice or establish that a less discriminatory alternative was unavailable.

This article is provided for general informational purposes and reflects the state of the law as of November 2026. It is not legal advice. Regulations in this area are changing quickly. Confirm current requirements and consult qualified counsel before making decisions for your organization.