The following is the Aug. 4, 2023, Congressional Research Service report, Artificial Intelligence: Overview, Recent Advances, and Considerations for the 118th Congress.
From the report
Artificial intelligence (AI)—a term generally thought of as computerized systems that work and react in ways commonly thought to require intelligence—can encompass a range of technologies, methodologies, and application areas, such as natural language processing, facial recognition, and robotics. The concept of AI has existed for decades, with the term first being coined in the 1950s, followed by alternating periods of much development and lulls in activity and progress.
A notable area of recent advancement has been in generative AI (GenAI), which refers to machine learning (ML) models developed through training on large volumes of data in order to generate content. Technological advancements in the underlying models since 2017, combined with the open availability of these tools to the public in late 2022, have led to widespread use. The underlying models for GenAI tools have been described as “general-purpose AI,” meaning they can be adapted to a wide range of downstream tasks. Such advancements, and the wide variety of applications for AI technologies, have renewed debates over appropriate uses and guardrails, including in the areas of health care, education, and national security.
AI technologies, including GenAI tools, have many potential benefits, such as accelerating and providing insights into data processing, augmenting human decisionmaking, and optimizing performance for complex systems and tasks. GenAI tools, for example, are increasingly capable of performing a broad range of tasks, such as text analysis, image generation, and speech recognition. However, AI systems may perpetuate or amplify biases in the datasets on which they are trained; may not yet be able to fully explain their decisionmaking; and often depend on such vast amounts of data and other resources that they are not widely accessible for research, development, and commercialization beyond a handful of technology companies.
Numerous federal laws on AI have been enacted over the past few Congresses, either as standalone legislation or as AI-focused provisions in broader acts. These include the expansive National Artificial Intelligence Initiative Act of 2020 (Division E of P.L. 116-283), which included the establishment of an American AI Initiative and direction for AI research, development, and evaluation activities at federal science agencies. Additional acts have directed certain agencies to undertake activities to guide AI programs and policies across the federal government (e.g., the AI in Government Act of 2020, P.L. 116-260; and the Advancing American AI Act, Subtitle B of P.L. 117-263). In the 117th Congress, at least 75 bills were introduced that either focused on AI and ML or had AI/ML-focused provisions. Six of those were enacted.
In the 118th Congress, as of June 2023, at least 40 bills had been introduced that either focused on AI/ML or contained AI/ML-focused provisions, and none has been enacted. Collectively, bills in the 118th Congress address a range of topics, including federal government oversight of AI; training for federal employees; disclosure of AI use; export controls; use-specific prohibitions; and support for the use of AI in particular sectors, such as cybersecurity, weather modeling, wildfire detection, precision agriculture, and airport safety.
A primary consideration under debate in the United States and internationally is whether and how to regulate AI technologies. The European Union’s draft Artificial Intelligence Act would broadly take a risk-based approach to regulatory requirements and prohibitions for certain uses. In the United States, previously introduced legislation has sought to require impact assessments and reporting for automated decision systems—including but not limited to AI systems—in critical areas (e.g., health care, employment, and criminal justice). Other perspectives on AI regulation have suggested a sector-specific approach with interagency coordination. A general concern for Congress might be how to approach AI regulatory efforts in a way that balances support for innovation and beneficial uses while minimizing current and future harms.
Additional considerations for the 118th Congress might include whether the current federal government mechanisms are sufficient for AI oversight and policymaking, the role of the federal government in supporting AI research and development, the potential impact of AI technologies on the workforce, disclosure of AI use, testing and validation of AI systems, and potential ways to support the development of trustworthy and responsible AI.
Download the document here.