By: Harrison H.B. No. 3808
A BILL TO BE ENTITLED
AN ACT
relating to the creation of the artificial intelligence advisory
council and the establishment of the artificial intelligence
learning laboratory.
BE IT ENACTED BY THE LEGISLATURE OF THE STATE OF TEXAS:
SECTION 1. Chapter 2054, Government Code, is amended by
adding Subchapter S to read as follows:
SUBCHAPTER S. ARTIFICIAL INTELLIGENCE ADVISORY COUNCIL
Sec. 2054.621. DEFINITIONS. In this subchapter:
(1) "Algorithm" means a computerized procedure
consisting of a set of steps used to accomplish a determined task.
(2) "Artificial intelligence systems" means systems
capable of:
(A) perceiving an environment through data
acquisition and processing and interpreting the derived
information to take an action or actions or to imitate intelligent
behavior given a specific goal; and
(B) learning and adapting behavior by analyzing
how the environment is affected by prior actions.
(3) "Automated decision system" means an algorithm,
including an algorithm incorporating machine learning or other
artificial intelligence techniques, that uses data-based analytics
to make or support governmental decisions, judgments, or
conclusions.
(4) "Automated final decision system" means an
automated decision system that makes final decisions, judgments, or
conclusions without human intervention.
(5) "Automated support decision system" means an
automated decision system that provides information to inform the
final decision, judgment, or conclusion of a human decision maker.
(6) "Council" means the artificial intelligence
advisory council established under this subchapter.
Sec. 2054.622. ARTIFICIAL INTELLIGENCE ADVISORY COUNCIL.
(a) The council is composed of the following seven members:
(1) one member of the house of representatives
appointed by the speaker of the house of representatives;
(2) one member of the senate appointed by the
lieutenant governor;
(3) the executive director or the executive director's
designee; and
(4) the following four members appointed by the
governor:
(A) an expert on ethics;
(B) an expert on artificial intelligence
systems;
(C) an expert on law enforcement usage of
artificial intelligence systems; and
(D) an expert on constitutional and legal rights.
(b) The council members appointed under Subsections (a)(1)
and (2) shall serve as co-chairs of the council.
(c) A member of the council is not entitled to compensation
or reimbursement for expenses.
(d) The department shall provide administrative support for
the council.
(e) The council shall meet at the call of the co-chairs. The
council may meet in person or by telephone conference call,
videoconference, or another similar telecommunication method.
Notwithstanding Chapter 551 or any other law, a meeting held by
telephone conference call, videoconference, or another similar
telecommunication method is subject to the requirements of Sections
551.125(c), (d), (e), and (f).
(f) The council shall study and monitor artificial
intelligence systems developed, employed, or procured by state
agencies. In carrying out its duties under this section, the
council shall:
(1) assess the need for a state code of ethics for
artificial intelligence systems in state government;
(2) review automated decision systems inventory
reports submitted by state agencies under Section 2054.623,
including a review of:
(A) the effect of the automated decision systems
on the constitutional or legal rights, duties, or privileges of the
residents of this state; and
(B) the potential benefits, liabilities, or
risks that this state could incur as a result of implementing the
automated decision systems; and
(3) recommend administrative actions that state
agencies may take without further legislative authorization.
(g) Not later than December 1, 2026, the council shall
submit a report to the governor and the legislature that includes:
(1) a summary of the council's findings after
reviewing:
(A) the automated decision systems inventory
reports submitted under Section 2054.623; and
(B) the results of the artificial intelligence
learning laboratory under Section 2054.624;
(2) a summary of the recommendations of any relevant
national bodies on artificial intelligence systems in state
government;
(3) an assessment of the impact of using artificial
intelligence systems in state government on the liberty, finances,
livelihood, and privacy interests of the residents of this state;
(4) recommendations of the least stringent policies
necessary to:
(A) protect the privacy and interests of the
residents of this state from any diminution caused by employment of
artificial intelligence systems by state government;
(B) ensure that the residents of this state are
free from unfair discrimination caused or compounded by the
employment of artificial intelligence systems in state government;
and
(C) promote workforce knowledge of artificial
intelligence technology and the development of ethical artificial
intelligence systems in state government; and
(5) any other information that the council considers
relevant.
(h) Section 2110.008 does not apply to the council.
Sec. 2054.623. AUTOMATED DECISION SYSTEMS INVENTORY
REPORT. (a) Not later than July 1, 2026, each agency in the
executive and legislative branches of state government, using money
appropriated to the agency by this state, shall submit an inventory
report of all automated decision systems that are being developed,
employed, or procured by the agency. For each automated decision
system, the inventory report must include a description of:
(1) the name and vendor of the automated decision
system, if any;
(2) the automated decision system's general
capabilities, including:
(A) reasonably foreseeable capabilities outside
the scope of the agency's proposed use; and
(B) whether the automated decision system is used
or may be used for independent decision-making powers and the
impact of those decisions on the residents of this state;
(3) the types of data inputs that the technology uses;
(4) how the data described by Subdivision (3) is
generated, collected, and processed;
(5) the types of data the automated decision system is
reasonably likely to generate;
(6) whether the automated decision system has been
tested by an independent third party, has a known bias, or is
untested for bias;
(7) the purpose and proposed use of the automated
decision system, including:
(A) the decisions the automated decision system
will be used to make or support;
(B) whether the automated decision system is an
automated final decision system or an automated support decision
system; and
(C) the automated decision system's intended
benefits, including any data or research relevant to the outcome of
those results;
(8) how automated decision system data is securely
stored and processed and whether the agency intends to share access
to the automated decision system or data from that automated
decision system with any other entity, and why; and
(9) the information technology fiscal impacts of the
automated decision system, including:
(A) initial acquisition costs and ongoing
operating costs, such as maintenance, licensing, personnel, legal
compliance, use auditing, data retention, and security costs;
(B) any cost savings that would be achieved
through the use of the technology; and
(C) any current or potential sources of funding,
including any subsidies or free products being offered by vendors
or governmental entities.
(b) Not later than March 1, 2026, the council, in
consultation with the department, shall prescribe the form,
contents, and manner of submission of the automated decision
systems inventory report required under this section.
(c) Each agency shall submit the report required under this
section to the:
(1) department;
(2) council; and
(3) standing committees of the senate and house of
representatives with primary jurisdiction over state agency
information technology.
Sec. 2054.624. ARTIFICIAL INTELLIGENCE LEARNING
LABORATORY. (a) In this section:
(1) "Learning laboratory" means the artificial
intelligence learning laboratory established under this section.
(2) "Participant" means an entity selected by the
council to participate in the artificial intelligence learning
laboratory established under this section, including a state
agency, a business entity, a nonprofit organization, or an open
source project.
(3) "Regulatory mitigation agreement" means an
agreement between the council and a participant for the purpose of
permitting the participant to use and test artificial intelligence
systems with limited legal and regulatory boundaries.
(4) "State agency" means, notwithstanding Section
2054.003, an agency in the legislative or executive branch of state
government.
(b) The council shall, with assistance from the department,
establish the artificial intelligence learning laboratory. The
council shall operate the learning laboratory to:
(1) evaluate the benefits, risks, and impact of
regulations on artificial intelligence;
(2) encourage the development of artificial
intelligence in state agencies; and
(3) produce recommendations for the least stringent
legislation on or regulation of artificial intelligence.
(c) The council shall set, and update as necessary, a
learning agenda for the learning laboratory to establish specific
areas of artificial intelligence policy to study. The learning
agenda must outline the parameters for the type of artificial
intelligence systems to be tested by a participant. In setting the
learning agenda, the council may consult state agencies, industry
leaders, and other key stakeholders the council considers
appropriate.
(d) The council shall establish an application process and
criteria to select participants for the learning laboratory. The
council shall, in selecting participants, consider:
(1) the relevance and utility of the applicant's
artificial intelligence technology to the learning agenda;
(2) the applicant's expertise and knowledge specific
to the learning agenda;
(3) the applicant's plan to monitor and minimize use
when testing the artificial intelligence systems;
(4) whether the applicant is eligible for regulatory
mitigation as described by Section 2054.626; and
(5) any other factor the council determines relevant.
(e) In collaboration with each participant, the council
shall establish benchmarks and assess outcomes of participation in
the learning laboratory.
(f) An applicant approved for participation must enter into
a regulatory mitigation agreement under Section 2054.625.
(g) A participant of the learning laboratory shall conduct
experiments using artificial intelligence systems, automated
decision systems, automated final decision systems, or automated
support decision systems that fall within the scope of the learning
agenda.
(h) A participant shall retain all records and provide
requested information and findings to the council. A participant
shall submit a report to the council, including:
(1) a summary of the proposed artificial intelligence
systems;
(2) the results from operating the artificial
intelligence systems; and
(3) any other information the council determines as
necessary and relevant.
(i) The council shall submit to the legislature, the
standing committees of each house with jurisdiction over artificial
intelligence systems, and the governor a report containing the
results of the learning laboratory findings and suggestions of the
least stringent legal and regulatory policies in relation to the
artificial intelligence systems studied by the learning
laboratory.
(j) The department may adopt rules to implement this
section.
Sec. 2054.625. REGULATORY MITIGATION AGREEMENTS. (a) A
participant in the learning laboratory shall enter into an
agreement as provided by this section.
(b) The agreement must provide:
(1) limitations on the artificial intelligence
systems to be tested, including:
(A) the number and types of users;
(B) geographic limitations; and
(C) any other limitation the department
considers advisable;
(2) procedures to protect the health and safety of
this state during the testing period;
(3) that the participant does not have any expectation
of a property right or license to participate in the learning
laboratory;
(4) that committing certain violations of law or a
violation of the agreement, as determined by the department, is
grounds for removal from the learning laboratory;
(5) that participation in the learning laboratory does
not constitute an endorsement by the state and is not a factor in
any criminal or civil proceeding;
(6) that the state is not responsible for any claims,
liabilities, damages, losses, or expenses arising out of the
participant's involvement in the learning laboratory; and
(7) that the participant will:
(A) provide information to the department and any
other state agency the department considers necessary to
successfully test artificial intelligence systems under the
learning laboratory program; and
(B) report to the department under the terms
acceptable to the department.
(c) An agreement under this section may be entered into by a
participant, the department, and other state agencies the
department considers necessary for regulatory mitigation as
described by Subsection (d). The department shall consult with
other state agencies on appropriate terms of regulatory mitigation
to provide under Subsection (d).
(d) An agreement under this section may provide for the
following regulatory mitigation measures:
(1) circumstances under which a participant:
(A) is or is not required to provide restitution
to a user of the artificial intelligence system under testing;
(B) may have an opportunity to correct any
violations of law that occur from using the system described by this
section before penalties are imposed by the state; and
(C) may be entitled to reduced civil penalties
during the period in which the agreement is in effect; and
(2) any other measures to address risks that the
department considers advisable.
(e) An agreement under this section has a term of one year
and may be extended on request by the participant for an additional
year, if the request is submitted not later than 30 days before the
date that the agreement's term ends. The department shall approve
or deny the request before the date that the agreement's term ends.
Sec. 2054.626. ELIGIBILITY. An applicant for participation
in the learning laboratory is eligible for regulatory mitigation as
described by Section 2054.625 if:
(1) the applicant has the technical expertise and
capability to responsibly deploy and test the proposed artificial
intelligence systems;
(2) the applicant has sufficient financial resources
to meet obligations during the testing period;
(3) the proposed artificial intelligence systems
provide substantial consumer benefits that may outweigh identified
risks from permitting regulatory mitigation;
(4) the applicant has an effective plan to monitor and
minimize identified risks from testing; and
(5) the scale, scope, and duration of proposed testing
is appropriately limited based on risk assessments.
Sec. 2054.627. COUNCIL ABOLISHED; EXPIRATION OF
SUBCHAPTER. The council is abolished and this subchapter expires
September 1, 2029.
SECTION 2. (a) As soon as practicable after the effective
date of this Act but not later than October 1, 2025, the appropriate
appointing authorities shall appoint the members to the artificial
intelligence advisory council as required by Section 2054.622,
Government Code, as added by this Act.
(b) Not later than November 1, 2025, the advisory council
established under Subchapter S, Chapter 2054, Government Code, as
added by this Act, shall hold its initial meeting.
(c) Not later than December 1, 2025, the advisory council
established under Subchapter S, Chapter 2054, Government Code, as
added by this Act, shall establish the artificial intelligence
learning laboratory as required by Section 2054.624, Government
Code, as added by this Act.
SECTION 3. This Act takes effect immediately if it receives
a vote of two-thirds of all the members elected to each house, as
provided by Section 39, Article III, Texas Constitution. If this
Act does not receive the vote necessary for immediate effect, this
Act takes effect September 1, 2025.