(d) The term “AI red-teaming” means a structured testing effort to find flaws and vulnerabilities in an AI system, often in a controlled environment and in collaboration with developers of AI. Artificial Intelligence red-teaming is most often performed by dedicated “red teams” that adopt adversarial methods to identify flaws and vulnerabilities, such as harmful or discriminatory outputs from an AI system, unforeseen or undesirable system behaviors, limitations, or potential risks associated with the misuse of the system.
(e) The term “AI system” means any data system, software, hardware, application, tool, or utility that operates in whole or in part using AI.
(f) The term “commercially available information” means any information or data about an individual or group of individuals, including an individual’s or group of individuals’ device or location, that is made available or obtainable and sold, leased, or licensed to the general public or to governmental or non-governmental entities.
(g) The term “crime forecasting” means the use of analytical techniques to attempt to predict future crimes or crime-related information. It can include machine-generated predictions that use algorithms to analyze large volumes of data, as well as other forecasts that are generated without machines and based on statistics, such as historical crime statistics.
(h) The term “critical and emerging technologies” means those technologies listed in the February 2022 Critical and Emerging Technologies List Update issued by the National Science and Technology Council (NSTC), as amended by subsequent updates to the list issued by the NSTC.
(i) The term “critical infrastructure” has the meaning set forth in section 1016(e) of the USA PATRIOT Act of 2001, 42 U.S.C. 5195c(e).
(j) The term “differential-privacy guarantee” means protections that allow information about a group to be shared while provably limiting the improper access, use, or disclosure of personal information about particular entities.
(k) The term “dual-use foundation model” means an AI model that is trained on broad data; generally uses self-supervision; contains at least tens of billions of parameters; is applicable across a wide range of contexts; and that exhibits, or could be easily modified to exhibit, high levels of performance at tasks that pose a serious risk to security, national economic security, national public health or safety, or any combination of those matters, such as by:
- (i) substantially lowering the barrier of entry for non-experts to design, synthesize, acquire, or use chemical, biological, radiological, or nuclear (CBRN) weapons;
- (ii) enabling powerful offensive cyber operations through automated vulnerability discovery and exploitation against a wide range of potential targets of cyber attacks; or
- (iii) permitting the evasion of human control or oversight through means of deception or obfuscation.
Models meet this definition even if they are provided to end users with technical safeguards that attempt to prevent users from taking advantage of the relevant unsafe capabilities.
(l) The term “Federal law enforcement agency” has the meaning set forth in section 21(a) of Executive Order 14074 of May 25, 2022 (Advancing Effective, Accountable Policing and Criminal Justice Practices To Enhance Public Trust and Public Safety).
(m) The term “floating-point operation” means any mathematical operation or assignment involving floating-point numbers, which are a subset of the real numbers typically represented on computers by an integer of fixed precision scaled by an integer exponent of a fixed base.
(n) The term “foreign person” has the meaning set forth in section 5(c) of Executive Order 13984 of January 19, 2021 (Taking Additional Steps
