Review – HR 9466 Introduced – AI Standards

Last month, Rep Baird (R,IN) introduced HR 9466, the AI Development Practices Act of 2024. The bill would amend 15 USC 278h-1, Standards for artificial intelligence. It would require the National Institute of Standards and Technology (NIST) to develop “voluntary guidance for practices and guidelines relating to the development, release, and assessment of artificial intelligence systems”. No new funding would be authorized by this bill.

Moving Forward

Baird, and one of his six cosponsors {Rep Bonamici (D,OR)}, are members of the House Science, Space, and Technology Committee to which this bill was assigned for consideration. This means that there may be sufficient influence to see the bill considered in Committee. Especially with no new funding authorization, I see nothing in the bill that engender any organized opposition. I suspect that the bill would receive some level of bipartisan report, sufficient bipartisan support, in fact, to see the bill (if reported favorably by the SST Committee) considered by the full House under the suspension of the rules process.

Commentary

My interest in AI, at least as far as this blog is concerned, is the need for cybersecurity protections for what is, at base, a sophisticated, complex, self-correcting, computer program. This is briefly and ineffectively accomplished here by requiring the development of security benchmarks {(h)(1)(B)(iii)} and disclosure of security practices {(h)(1)(B)(vi)} (thus the need for the definition of security of the term “artificial intelligence red teaming”).

As the use of AI systems is being expanded in the area of chemical and biological process development and control, the need for such cybersecurity protections becomes more important. While all software deserves at least some level of such protections, process controls that could be used to cause catastrophic disruptions of facilities and local communities should arguably be required to have some minimal level protection against unauthorized manipulations.

 

For more information on the provisions of this bill, as well as my suggestions for some added cybersecurity related verbiage, see my article at CFSN Detailed Analysis - https://patrickcoyle.substack.com/p/hr-9466-introduced - subscription required.

Bills Introduced – 9-6-24

Yesterday, with the House meeting in pro forma session, there were 43 bills introduced. Five of those bills may receive additional attention in this blog:

HR 9458 To amend the Homeland Security Act of 2002 to enhance outreach for the Urban Area Security Initiative and the State Homeland Security Grant Program of the Department of Homeland Security, and for other purposes. Goldman, Daniel S. [Rep.-D-NY-10]

HR 9459 To amend the Homeland Security Act of 2002 to enable secure and trustworthy technology through other transaction contracting authority, and for other purposes. Guest, Michael [Rep.-R-MS-3] 

HR 9466 To direct the National Institute of Standards and Technology to catalog and evaluate emerging practices and norms for communicating certain characteristics of artificial intelligence systems, including relating to transparency, robustness, resilience, security, safety, and usability, and for other purposes. Baird, James R. [Rep.-R-IN-4] 

HR 9468 Making supplemental appropriations for the fiscal year ending September 30, 2024, and for other purposes.  Garcia, Mike [Rep.-R-CA-27

HR 9469 To amend the Homeland Security Act of 2002 to codify the Transportation Security Administration's responsibility relating to securing pipeline transportation and pipeline facilities against cybersecurity threats, acts of terrorism, and other nefarious acts that jeopardize the physical security or cybersecurity of pipelines, and for other purposes. Garcia, Robert [Rep.-D-CA-42]

I will be covering HR 9458, HR 9459, HR 9468, and HR 9469.

I will be watching HR 9466 for language and definitions that would specifically include AI applications designed to be used with operational technology systems.