Bill to Regulate the Use of AI Arrives in the Senate | Benefits of Generative ai | Bcg Generative ai pdf | Generative ai in Finance pdf | Turtles AI

US Senate considers PREPARED bill for artificial intelligence

Key points:
1. The PREPARED bill aims to regulate the use of AI in government agencies.
2. Agencies will be required to assess and address the risks associated with AI before purchasing and implementing tools.
3. The bill provides for the creation of governance structures and pilot programs for the procurement of AI technologies.
4. There are public transparency and reporting requirements on the government’s use of AI.

On July 31, 2024, the Department of Homeland Security and Governmental Affairs announced that Senators Gary Peters and Thom Tillis have introduced the Procurement to Advance Readiness for Enterprise-wide Deployment (PREPARED) for Artificial Intelligence Act. This bill, whose passage is currently under consideration in the U.S. Senate, aims to regulate the use of artificial intelligence (AI) within government agencies by ensuring that each application is carefully evaluated and managed to reduce associated risks.

Risk assessment and management
The PREPARED bill requires government agencies to conduct detailed risk assessments before purchasing and implementing AI tools. This includes ongoing testing and evaluation to identify and mitigate potential risks from the use of such technologies.

Governance structures and pilot programs
One of the central aspects of the bill is the creation of dedicated AI governance structures, which will be tasked with overseeing the implementation and use of these technologies in agencies. In addition, pilot programs will be launched to test and optimize processes for the procurement and integration of new AI technologies.

Transparency and accountability
The bill places great emphasis on transparency by requiring the government to publicly disclose information on the use of AI. This includes detailed reports on applications and achievements, thus ensuring continuous reporting that is accessible to all citizens.

Specific terms for AI procurement
The bill clearly defines several key terms, such as "adverse incident," "biometric data," "commercial technology," "distributor," and "risk." For instruments considered high-risk, it will be mandatory to incorporate specific contract terms and guidelines that include:
- Adverse incident reporting mechanisms.
- Reporting requirements by distributors in case of incidents.
- The ability for agencies to suspend the use of AI tools if risks become unacceptable.
- Requirements for data quality, provenance and ownership.
- Need for ongoing evaluation and testing of AI technologies.
- Requirement for distributors and developers to provide detailed documentation.

The implementation of these measures is expected within a year of the bill’s enactment, and is intended to create a safer and more transparent AI use environment within U.S. government agencies.