AI MODEL SAFETY

Illinois General Assembly - Full Text of SB3444

Select Language

×

The Illinois General Assembly offers the Google Translate™ service for visitor convenience. In no way should it be considered accurate as to the translation of any content herein.

Visitors of the Illinois General Assembly website are encouraged to use other translation services available on the internet.

The English language version is always the official and authoritative version of this website.

NOTE: To return to the original English language version, select the "Show Original" button on the Google Translate™ menu bar at the top of the window.

Choose Language

English

Afrikaans

Albanian

Arabic

Armenian

Azerbaijani

Basque

Bengali

Bosnian

Catalan

Croatian

Czech

Danish

Dutch

Esperanto

Estonian

Filipino

Finnish

French

Galician

Georgian

German

Greek

Gujarati

Haitian Creole

Hausa

Hawaiian

Hebrew

Hindi

Hungarian

Icelandic

Indonesian

Interlingua

Interlingue

Inuktitut

Irish

Italian

Japanese

Javanese

Kannada

Khmer

Korean

Latin

Latvian

Lithuanian

Luxembourgish

Macedonian

Malagasy

Malayalam

Maltese

Maori

Marathi

Myanmar

Nepali

Norwegian

Odia

Pashto

Punjabi

Romanian

Russian

Samoan

Sango

Sanskrit

Sardinian

Sindhi

Sinhala

Slovak

Slovenian

Somali

Southern Sotho

Spanish

Sundanese

Swahili

Swedish

Tamil

Telugu

Thai

Tigrinya

Tonga

Turkish

Ukrainian

Urdu

Vietnamese

Welsh

Xhosa

Yiddish

Yoruba

Zulu

Powered by
Translate

Close

Illinois General Assembly

Top Navigation Bar

Translate

Learn

Select General Assembly

Search the 104th General Assembly

Enter search terms for legislation, members, committees, or schedules.

ILGA.GOV

LEGISLATION & LAWS

Bills & Resolutions

Public Acts

Illinois Compiled Statutes

Illinois Constitution

Search Legislation

Glossary

Guide

Reports & Inquiry

Legislative Reports

Special Reports

FTP Site

Legislator Lookup

Capitol Complex Phone Numbers

Rules & Regulations

Illinois Register

Administrative Rules

Senate

Members

Schedules

Committees

Request for Remote Testimony

Journals

Transcripts

Rules

Audio/Video

FOIA Information

Senate Employment Opportunities

Media Guidelines

House

Members

Schedules

Committees

Submit testimony for House Committees

Journals

Transcripts

Rules

Audio/Video

FOIA Information

House Employment Opportunities

Log In

Mobile Top Bar

Search the 104th General Assembly

Enter keywords to search the Illinois General Assembly website.

Full Text of SB3444

Home

Legislation

Full Text

SB3444 - 104th General Assembly

Bill Status

Full Text

Votes

Witness Slips

Select Menu

Bill Status

Full Text

Votes

Witness Slips

Printer Friendly Version

Introduced

Printer Friendly Version

Introduced

Open PDF

104TH GENERAL ASSEMBLY
State of Illinois
2025 and 2026
SB3444

Introduced 2/4/2026, by Sen. Bill Cunningham

SYNOPSIS AS INTRODUCED:

New Act

Creates the Artificial Intelligence Safety Act. Provides that a
developer of a frontier artificial intelligence model shall not be held
liable for critical harms caused by the frontier model if the developer did
not intentionally or recklessly cause the critical harms and the developer
publishes a safety and security protocol and transparency report on its
website. Provides that a developer shall be deemed to have complied with
these requirements if the developer: (1) agrees to be bound by safety and
security requirements adopted by the European Union; or (2) enters into an
agreement with an agency of the federal government that satisfies
specified requirements. Sets forth requirements for safety and security
protocols and transparency reports. Provides that the Act shall no longer
apply if the federal government enacts a law or adopts regulations that
establish overlapping requirements for developers of frontier models.
LRB104 15770 SPS 28959 b

A BILL FOR

SB3444
LRB104 15770 SPS 28959 b
1

AN ACT concerning business.

2

Be it enacted by the People of the State of Illinois,
3
represented in the General Assembly:

4

Section 1.
Short title.
This Act may be cited as the
5
Artificial Intelligence Safety Act.

6

Section 5.
Definitions.
In this Act:
7

"Artificial intelligence model" means an engineered or
8
machine-based component of a system that varies in its level
9
of autonomy and that can, for explicit or implicit objectives,
10
infer from the input it receives how to generate outputs that
11
can influence physical or virtual environments.
12

"Critical harm" means the death or serious injury of 100
13
or more people or at least $1,000,000,000 of damages to rights
14
in property caused or materially enabled by a frontier model,
15
through either:
16

(1) the creation or use of a chemical, biological,
17

radiological, or nuclear weapon; or
18

(2) engaging in conduct that:
19

(A) acts with no meaningful human intervention;
20

and
21

(B) would, if committed by a human, constitute a
22

criminal offense that requires intent, recklessness,
23

or negligence, or the solicitation or aiding and

SB3444
- 2 -
LRB104 15770 SPS 28959 b
1

abetting of such a crime.
2

"Developer" means a person or organization that has
3
trained, or initiated the training of, at least one frontier
4
model.
5

"Frontier model" means an artificial intelligence model
6
that: (1) is trained using greater than 10^26 computational
7
operations, such as integer or floating-point operations; or
8
(2) has a compute cost that exceeds $100,000,000.

9

Section 10.
Frontier model transparency.
10

(a) A developer shall not be held liable for critical
11
harms if the developer did not intentionally or recklessly
12
cause the critical harms and the developer:
13

(1) published a safety and security protocol on its
14

website that satisfies the requirements of Section 15 and
15

adhered to that safety and security protocol prior to the
16

release of the frontier model;
17

(2) published a transparency report on its website at
18

the time of the frontier model's release that satisfies
19

the requirements of Section 20.
20

The requirements of paragraphs (1) and (2) do not apply if
21
the developer does not reasonably foresee any material
22
difference between the frontier model's capabilities or risks
23
of critical harm and a frontier model that was previously
24
evaluated by the developer in a manner substantially similar
25
to this Act.

SB3444
- 3 -
LRB104 15770 SPS 28959 b
1

(b) If a developer has produced a safety and security
2
protocol in a manner substantially similar to this Act, then
3
the developer shall be deemed to have complied with subsection
4
(a) if the developer:
5

(1) agrees to be bound by the safety and security
6

requirements adopted under Article 56 of the European
7

Union's Artificial Intelligence Act; or
8

(2) enters into an agreement with an agency of the
9

federal government that:
10

(A) enables the agency to access the developer's
11

frontier models for the purposes of conducting
12

research and evaluations;
13

(B) facilitates the evaluation of frontier models,
14

such as an assessment of cyber and biological risks;
15

and
16

(C) allows for the federal government to release
17

information related to evaluations of frontier models
18

that have been publicly released.
19

(c) A developer that enters into an agreement with an
20
agency of the federal government under paragraph (2) of
21
subsection (b) shall file a certification with the Attorney
22
General, in a manner prescribed by the Attorney General,
23
attesting that the developer has entered into an agreement
24
that satisfies the requirements of paragraph (2) of subsection
25
(b). It is a violation of this Act to make a materially false
26
or misleading statement, or to omit a material fact, in any

SB3444
- 4 -
LRB104 15770 SPS 28959 b
1
certification submitted under this subsection.

2

Section 15.
Requirements for safety and security
3
protocols.
4

(a) In order to satisfy the requirements of paragraph (1)
5
of subsection (a) of Section 10, a developer shall create a
6
safety and security protocol that documents a developer's
7
technical and organizational protocols to manage, assess, and
8
mitigate risk of critical harm arising from the use and
9
distribution of its frontier models. It shall include a
10
high-level summary of:
11

(1) testing procedures that the developer uses to
12

assess risk of a reasonably foreseeable critical harms
13

arising from the deployment of a frontier model;
14

(2) thresholds used by the developer to identify and
15

assess whether a frontier model poses a critical risk,
16

including:
17

(A) how the developer will assess whether a
18

threshold has been attained, which may include
19

multiple tiered thresholds; and
20

(B) an illustrative summary of the potential
21

actions the developer may take if each threshold is
22

attained;
23

(3) the mitigations that a developer takes to mitigate
24

reasonably foreseeable critical harms and how the
25

developer assesses the effectiveness of those mitigations;

SB3444
- 5 -
LRB104 15770 SPS 28959 b
1

(4) whether and how a developer will use third parties
2

to assess risk of critical harm and capabilities or the
3

effectiveness of mitigations of critical harms;
4

(5) the developer's cybersecurity practices relating
5

to frontier models and how the developer secures
6

unreleased frontier model weights from unauthorized
7

modification or transfer by internal or external parties;
8

(6) to the extent that the frontier model is deployed
9

by the developer, a description of how the developer will
10

monitor for critical harm and how a developer would
11

respond; and
12

(7) the process by which the developer determines when
13

a frontier model presents material new risks such that it
14

should be subject to additional assessments.
15

(b) A developer may make appropriate redactions to the
16
information required under subsection (a), as necessary, to
17
maintain model security and integrity, trade secrets, and
18
proprietary information.

19

Section 20.
Requirements for transparency reports.
20

(a) In order to satisfy the requirements of paragraph (2)
21
of subsection (a) of Section 10, a developer shall create a
22
transparency report that:
23

(1) identifies the frontier model that is the subject
24

of the transparency report; and
25

(2) provides a summary of the results of assessments

SB3444
- 6 -
LRB104 15770 SPS 28959 b
1

conducted in accordance with the developer's safety and
2

security protocol and the steps taken to address any
3

identified risks.
4

(b) A developer may make appropriate redactions to the
5
information required under subsection (a), as necessary, to
6
maintain model security and integrity, trade secrets, and
7
proprietary information.

8

Section 25.
Application of the Act.
9

(a) This Act shall no longer apply if the federal
10
government enacts a law or adopts regulations that establish
11
overlapping requirements for developers of frontier models.
12

(b) A developer shall be deemed to be in compliance with
13
this Act if the developer is in compliance with another
14
state's substantially similar frontier model safety and
15
security framework.

Footer

Disclaimer

This site is maintained for the Illinois General Assembly by the
Legislative Information System, 705 Stratton Building, Springfield, Illinois 62706.

Contact ILGA Webmaster

ILGA.gov uses cookies to ensure you get the best experience on our website. By continuing to browse ILGA.gov you consent to our use of cookies.
Read About Cookies

ILGA.GOV

2026 ILGA.gov | All Rights Reserved |
ADA

|
Disclaimers
|
Learn

This site is maintained for the Illinois General Assembly by the
Legislative Information System, 705 Stratton Building, Springfield, Illinois 62706.
Contact ILGA Webmaster

ILGA.gov uses cookies to ensure you get the best experience on our website. By continuing to browse ILGA.gov you consent to our use of cookies.
Read About Cookies

ILGA.GOV

2026 ILGA.gov | All Rights Reserved |
ADA

|
Disclaimers
|
Learn