GeoAI and the Law Newsletter
Keeping geospatial professionals informed on the legal and policy issues that will impact GeoAI.
Summary of Recent Developments in GeoAI and the Law
This week’s edition of GeoAI and the Law touches upon a wide range of items, including AI audits, regulation of financial institutions for use of AI for certain transaction and a report that Brazil has imposed restrictions on the use of personal information of its citizens in training data. The Deep Dive looks at how regulating the use of Ai by financial institutions could impact geospatial companies.
Recommended Reading
INTELLIGENCE, HUMAN RIGHTS, DEMOCRACY, AND THE RULE OF LAW: REGULATION OF AI IN BROAD STROKES
This article discusses the Council of Europe's Framework Convention on Artificial Intelligence and Human Rights, Democracy, and the Rule of Law (the “Convention”), adopted in May 2024. The article discusses the Convention's potential to influence global AI governance and promote interoperability between different regulatory frameworks.
The European Data Protection Board (EDPB) is an independent European body. Its mission is to ensure that the General Data Protection Regulation (GDPR) is applied consistently, including in enforcement. The AI Audit tools are intended to help national Data Protection Authorities (DPAs) to inspect AI systems. It includes a checklist for performing an audit of an algorithm and proposes tools that would enhance transparency.
OCC, FDIC adopt final rule on automated valuation models (ABA Banking Journal)
The article discusses a final rule published by several regulators of financial institutions intended to satisfy a requirement in the Dodd-Frank Act to require quality control standards for Automated Valuation Models (AVMs) used in valuing real estate collateral that secures mortgage loans. According to the article, the final rule limits the scope of the regulation to credit decisions, including loan modifications, and covered securitization determinations. It will take effect one year following its publication in the Federal Register.
Department of Energy Generative Artificial Intelligence Reference Guide
The reference guide issued by the Department of Energy includes key concepts and opportunities for applying GenAI within the Department of Energy, as well as a discussion of key considerations, risks, best practices, and recommendations. It is intended to offer an overview of the specific risks, considerations, responsibilities, and recommendations that are associated with various organizational roles. It is meant to supplement, but not replace, the existing regulations surrounding GenAI.
Brazil data regulator bans Meta from mining data to train AI models (AP)
According to the article, “Brazil’s national data protection authority determined on Tuesday that Meta, the parent company of Instagram and Facebook, cannot use data originating in the country to train its artificial intelligence.”
The Deep Dive
As discussed above, several government agencies, including the Federal Reserve, the Federal Deposit Insurance Corporation (FDIC) and the Consumer Financial Protection Bureau (CFPB) issued a final rule in June relating to quality control standards for AVMs. (For purposes of the rule, AVM is defined as “any computerized model used by mortgage originators and secondary market issuers to determine the collateral worth of a mortgage secured by a consumer’s principal dwelling.”).
Under the rule, financial institutions that use AVMs to make credit decisions, such as loan modifications and certain securitization decisions, must adopt policies, practices, procedures and control systems to ensure that the AVMs adhere to quality control standards. These standards should “ensure a high level of confidence in the estimates produced” by the model” by:
(1) ensuring a high level of confidence in the estimates produced;
(2) protecting against the manipulation of data;
(3) avoiding conflicts of interest;
(4) requiring random sample testing and reviews; and
(5) complying with applicable nondiscrimination laws.
While the rule will only apply to financial institutions, it states that these institutions should “work with AVM developers and vendors to assist them with their compliance obligations under the rule, as they do with other third-party vendors in order to comply with relevant regulatory requirements.” These requirements will likely be passed down to geospatial companies that are assisting in the development of AVM models – either directly or indirectly – through contracts with terms that require geospatial companies adopt similar policies, practices, procedures and control systems.
Location data and other types of geospatial information are used by financial institutions in a variety of ways. With respect to determining a collateral’s worth, geospatial information is used to help determine and analyze property values. This helps financial institutions to “evaluate a borrower's assets” and assess credit risk. As a result, geospatial companies that sell geospatial information or GeoAI models to be used in determining property values to financial institutions will likely be asked to provide assurances about the level of confidence in their information and/or models and also provide random testing and reviews.
A recent Financial Times special report, the “Future of AI” notes of the growing of need for vendor-customer partnerships for AI systems “to be far more collaborative and long-lasting than in the past.” The report includes a quote from Reeves Givens of the Center for Democracy and Technology, that vendors and supplies “have a shared responsibility to get [supply chains] right.” This rule is an example of how regulators intend to foster such closer relationships.