Bill

Bill > HB1399


MD HB1399

MD HB1399
Consumer Protection - Consumer Reporting Agencies - Use of Algorithmic Systems


summary

Introduced
02/13/2026
In Committee
02/13/2026
Crossed Over
Passed
Dead

Introduced Session

2026 Regular Session

Bill Summary

Establishing requirements for consumer reporting agencies that use algorithmic systems to assemble or evaluate consumer credit information on consumers for the purpose of furnishing consumer reports to third parties; and requiring the Commissioner of Financial Regulation of the Maryland Department of Labor to establish certain assessment thresholds for algorithms, mandate regular training for human reviewers, and implement a certain whistleblower protection program.

AI Summary

This bill establishes new requirements for consumer reporting agencies, which are companies that collect and evaluate consumer credit information to provide consumer reports to third parties, when they use algorithmic systems, meaning automated processes, to make these evaluations. Specifically, these agencies must be able to explain their algorithmic evaluations in plain language, maintain a public registry of the algorithms they use, including their purpose and data sources, and meet strict error rate thresholds for both overall evaluations and discriminatory data based on protected characteristics, while also ensuring high accuracy for data inputs. They are also required to undergo quarterly audits by independent third parties to assess harmful bias and submit these audits and annual performance reports to the Commissioner of Financial Regulation, who oversees financial regulations in Maryland. Furthermore, these agencies must implement a data governance framework to ensure data accuracy and relevance, track data lineage, and meet minimum data set size requirements for statistical significance in their evaluations. Before a final decision is made, all automated evaluations must be reviewed by a human within 24 hours, with an expedited review process available within 48 hours of a request, and at least 10% of all algorithmic evaluations must undergo random human review. The bill also mandates that agencies designate staff for compliance, implement systems to improve algorithmic performance based on reviews, offer non-algorithmic assessment options for consumers who opt out of automated decisions, and maintain a contingency plan for system failures or data breaches. Consumers must be notified when artificial intelligence is used in an evaluation, and any artificially generated outputs, such as synthetic audio or text, must be detectable as such. The Commissioner will establish annual assessment thresholds for algorithms, mandate regular training for human reviewers on algorithm functionality and potential bias, and implement a whistleblower protection program for employees who report algorithmic misuse or noncompliance, with the Commissioner also empowered to adopt further regulations.

Committee Categories

Business and Industry

Sponsors (1)

Last Action

House Economic Matters Hearing (13:00:00 3/10/2026 ) (on 03/10/2026)

bill text


bill summary

Loading...

bill summary

Loading...
Loading...