AI training data

AN ACT Relating to artificial intelligence training data; and 1
adding a new chapter to Title 19 RCW. 2
BE IT ENACTED BY THE LEGISLATURE OF THE STATE OF WASHINGTON:3
NEW SECTION. Sec. 1. The definitions in this section apply 4
throughout this chapter unless the context clearly requires 5
otherwise.6
(1) "Aggregate consumer information" means information that 7
relates to a group or category of consumers, from which individual 8
consumer identities have been removed, that is not linked or 9
reasonably linkable to any consumer or household, including via a 10
device. "Aggregate consumer information" does not mean one or more 11
individual consumer records that have been deidentified.12
(2) "Artificial intelligence" means the use of machine learning 13
and related technologies that use data to train statistical models 14
for the purpose of enabling computer systems to perform tasks 15
normally associated with human intelligence or perception, such as 16
computer vision, speech, or natural language processing, and content 17
generation. 18
(3) "Developer" means a person, partnership, or corporation 19
primarily engaged in developing or substantially modifying a 20
generative artificial intelligence system intended for commercial 21
H-2895.1
HOUSE BILL 2503
State of Washington 69th Legislature 2026 Regular Session
By Representatives Shavers, Duerr, Parshley, Kloba, Hill, Pollet, and
Ramel
Read first time 01/15/26. Referred to Committee on Technology,
Economic Development, & Veterans.
p. 1 HB 2503
distribution or public use. For the purposes of this subsection, 1
"public use" does not include use by an affiliate as defined in RCW 2
19.146.010. This definition excludes individuals or entities that 3
develop artificial intelligence systems solely for internal use or 4
research purposes or those that utilize third-party artificial 5
intelligence systems via application programming interface without 6
substantial modification. This definition also excludes public 7
entities and tribal nations. 8
(4) "Generative artificial intelligence" means an artificial 9
intelligence system that generates novel data or content based on a 10
foundation model. 11
(5) "Security and integrity" means the ability of:12
(a) Networks or information systems to detect security incidents 13
that compromise the availability, authenticity, integrity, and 14
confidentiality of stored or transmitted personal information;15
(b) Developers, users, or businesses to detect security 16
incidents, resist malicious, deceptive, fraudulent, or illegal 17
actions and to help prosecute those responsible for those actions; 18
and 19
(c) Developers, users, or businesses to ensure the physical 20
safety of natural persons. 21
(6) "Substantial modification" means a new version, new release, 22
or substantial update from the developer, or an intentional and 23
deliberate change to a generative artificial intelligence system by a 24
deployer, that materially changes its functionality or performance in 25
a manner that was not reasonably foreseeable to the developer at the 26
time the artificial intelligence system was made publicly available 27
by the developer. 28
(7) "Synthetic data generation" means a process in which original 29
data are used to create artificial data that have some of the 30
statistical characteristics of the original data. 31
(8) "Train a generative artificial intelligence system" includes 32
testing, validating, or fine tuning by the developer of the 33
generative artificial intelligence system. 34
NEW SECTION. Sec. 2. (1) On or before January 1, 2027, and 35
before each time thereafter that a generative artificial intelligence 36
system, or a substantial modification to a generative artificial 37
intelligence system, released on or after January 1, 2022, is made 38
publicly available to Washingtonians for use, regardless of whether 39
p. 2 HB 2503
the terms of that use include compensation, the developer of the 1
system shall post on the developer's internet website documentation 2
regarding the data used by the developer to train the generative 3
artificial intelligence system including, but not limited to, a high-4
level summary of the datasets used in the development of the 5
generative artificial intelligence system including, but not limited 6
to: 7
(a) The sources of the datasets; 8
(b) A general description of how the datasets further the 9
intended purpose of the generative artificial intelligence system;10
(c) The number of data points included in the datasets, which may 11
be in general ranges, and with estimated figures for dynamic 12
datasets; 13
(d) A high-level description of the types of data points within 14
the datasets; 15
(e) Whether the datasets include any data protected by copyright, 16
trademark, or patent, or whether the datasets are entirely in the 17
public domain; 18
(f) Whether the datasets were purchased or licensed by the 19
developer; 20
(g) Whether the datasets include personal information, as defined 21
in RCW 19.373.010; 22
(h) Whether the datasets include aggregate consumer information;23
(i) Whether there was any cleaning, processing, or other 24
modification to the datasets by the developer, including the intended 25
purpose of those efforts in relation to the generative artificial 26
intelligence system; 27
(j) The dates the datasets were first trained or the date of the 28
last significant update to the datasets during the development of the 29
generative artificial intelligence system; and 30
(k) Whether the generative artificial intelligence system used or 31
continuously uses synthetic data generation in its development. A 32
developer may include a description of the functional need or desired 33
purpose of the synthetic data in relation to the intended purpose of 34
the system. 35
(2) A developer is not required to post documentation regarding 36
the data used to train a generative artificial intelligence system 37
for any of the following: 38
(a) A generative artificial intelligence system whose sole 39
purpose is to help ensure security and integrity; 40
p. 3 HB 2503
(b) A generative artificial intelligence system whose sole 1
purpose is the operation of aircraft in the national airspace; and2
(c) A generative artificial intelligence system developed for 3
national security, military, or defense purposes that is made 4
available only to a federal entity. 5
(3) The requirements in this chapter do not apply to generative 6
artificial intelligence systems that are subject to applicable 7
requirements under the federal food, drug, and cosmetic act, 21 8
U.S.C. Sec. 301 et seq., as amended. 9
NEW SECTION. Sec. 3. Section 2 of this act shall be construed 10
to require developers to comply based on the generally acknowledged 11
state of the art, which may include, but is not limited to, guidance 12
issued by the national institute of standards and technology, without 13
compromising their own intellectual property rights or trade secrets.14
NEW SECTION. Sec. 4. The legislature finds that the practices 15
covered by this chapter are matters vitally affecting the public 16
interest for the purpose of applying the consumer protection act, 17
chapter 19.86 RCW. A violation of this chapter is not reasonable in 18
relation to the development and preservation of business and is an 19
unfair or deceptive act in trade or commerce and an unfair method of 20
competition for purposes of applying the consumer protection act, 21
chapter 19.86 RCW.22
NEW SECTION. Sec. 5. Sections 1 through 4 of this act 23
constitute a new chapter in Title 19 RCW.24
--- END ---
p. 4 HB 2503