| AI Summary |
This bill modifies Colorado’s 2024 AI law (SB24-205) by narrowing its scope and easing compliance burdens, especially for small developers and less impactful AI systems.
It redefines “algorithmic discrimination” to mean violations of existing anti-discrimination laws only. Developers are exempt from certain obligations if they release models with open weights or meet size-based criteria (e.g., <$10M in funding, <$5M revenue, <5 years in operation).
“High-risk” AI systems are now limited to those that directly or primarily make consequential decisions. Tools that play a minor or assistive role are excluded.
The bill eliminates duties for developers and deployers to prevent foreseeable harms or notify the Attorney General of risks. It also loosens transparency rules for small deployers or those using AI only for external hiring.
New requirements expand impact assessments to include risks related to accessibility, deceptive practices, labor law, and privacy. Consumer appeal rights apply only to non-time-limited, non-competitive decisions. Companies that withhold disclosures must now justify doing so.
Attorney General enforcement begins January 1, 2027, giving industry time to adjust. Supporters say the changes balance safety and innovation; critics warn they reduce protections for those affected by high-risk AI decisions.
|
| Summary |
In 2024, the general assembly enacted Senate Bill 24-205, which
created consumer protections in interactions with artificial intelligence
systems (provisions). The bill amends these provisions by:
•
Redefining algorithmic discrimination to mean the use of
an artificial intelligence system that results in a violation of
any applicable local, state, or federal anti-discrimination
law;
•
Creating an exception to the definition of developer of an
artificial intelligence system (developer) if a person offers
the artificial intelligence system with open model weights
or if the person meets specified conditions regarding the
artificial intelligence system;
•
Exempting specified technologies that do not make, or are
not a substantial factor in making, a consequential decision
from the definition of high-risk artificial intelligence
system;
•
Eliminating the duty of a developer or deployer of a
high-risk artificial intelligence system (deployer) to use
reasonable care to protect consumers from any known or
reasonably foreseeable risks of algorithmic discrimination;
•
Eliminating the requirement that a developer or deployer
notify the attorney general of any known or reasonably
foreseeable risks of algorithmic discrimination arising from
the intended uses of the high-risk artificial intelligence
system;
•
Exempting a developer from specified disclosure
requirements if the developer has received less than
$10,000,000 from third-party investors, has annual
revenues of less than $5,000,000, and has been actively
operating and generating revenue for less than 5 years and
sells, distributes, or otherwise makes available to deployers
high-risk artificial intelligence systems that do not exceed
specified limits on the number of consequential decisions
made by the systems;
•
Requiring a deployer to include in an impact assessment
whether the system poses any known or reasonably
foreseeable risks of limiting accessibility for certain
individuals, an unfair or deceptive trade practice, a
violation of state or federal labor laws, or a violation of the
Colorado Privacy Act;
•
Requiring a deployer to provide additional information to
a consumer if the high-risk artificial intelligence system
makes, or is a substantial factor in making, a consequential
decision concerning the consumer;
•
Amending provisions regarding a consumer's right to
appeal an adverse consequential decision concerning the
consumer so that the provisions apply only to an adverse
consequential decision that is not a time-limited decision or
a competitive decision;
•
Clarifying the meaning of adverse when referring to a
consequential decision;
•
Broadening an exemption for a deployer from specified
disclosure requirements based on the deployer's number of
full-time equivalent employees;
•
Exempting a deployer from specified requirements if the
deployer uses the high-risk artificial intelligence system
solely relating to the recruitment, sourcing, or hiring of
external candidates for employment, meets specified
disclosure requirements, and does not employ more than
specified limits on the number of full-time equivalent
employees;
•
Applying specified requirements only to high-risk artificial
intelligence systems that make, or are the principal basis in
making, consequential decisions;
•
Requiring a developer or deployer that withholds
information otherwise subject to disclosure to provide
specified information regarding the disclosure; and
•
Requiring that the attorney general's authority to investigate
and enforce violations of the provisions begins on January
1, 2027.
|
