Getting set to tackle the indiscriminate use of AI, Colorado’s AI discrimination law is a landmark moment in the work being done in favor of AI regulations. The use of artificial intelligence has brought major changes into the work setting and while not all of them are bad, they are carriers of discrimination to a degree. The Colorado AI regulations target deployers and developers of “high-risk artificial intelligence systems” to ensure consumers are protected from any discrimination the algorithm might dole out. Colorado’s AI discrimination bill goes into effect on Feb 1, 2026, but employers and other agencies will have to begin planning ahead in adherence to these regulations.
For example, employers who actively use AI for hiring need to ensure that there are no algorithm issues that are denying candidates an opportunity on any discriminatory basis. Even if the AI is not one they have created themselves, as deployers of the AI, they need to be well-informed about its functionality and any changes that are made to it.
Image: Freepik
Colorado’s AI Discrimination Law—What is Algorithmic Discrimination?
Colorado’s AI discrimination bill, Senate Bill 24-205, was signed into effect on Friday by Governor Jared Polis, and it focuses on the prevention of algorithmic discrimination that may prevent members of a specific community from availing of an opportunity. It pertains to “high-risk artificial intelligence systems” which refers to any AI system that contributes to making consequential decisions, especially those that function without human review. It does not include AI systems that detect patterns or perform minor procedural tasks, which is reasonable as those have fewer opportunities for discrimination.
Additionally, the Colorado AI bias law applies to deployers—those who have business in the state and make use of these high-risk AI systems—and developers—those who develop or modify these AI systems.
“Algorithmic discrimination” means any condition in which the use of an artificial intelligence system results in an unlawful differential treatment or impact that disfavors an individual or group of individuals on the basis of their actual or perceived age, color, disability, ethnicity, genetic information, limited proficiency in the English language, national origin, race, religion, reproductive health, sex, veteran status, or other classification protected under the laws of this state or federal law.
—Colorado’s AI Discrimination Law
From the explanation of what the law means by algorithmic discrimination, it is clear that the governing body aims to protect individuals from being impacted by bias in AI. The law applies to companies and organizations using high-risk AI systems, especially those involved in hiring or employment decisions, and it is up to them to check their systems regularly to protect those who come in contact with their services from unfair treatment.
What Does Colorado’s AI Discrimination Bill Mean for Deployers?
In adherence to the Colorado AI law, from 2026, deployers of AI systems that meet the guidelines mentioned in the bill will have to take “reasonable care” to protect their consumers from being affected by any kind of algorithmic discrimination. These deployers need to ask the AI systems for any foreseeable risks that might occur in the future to cover as much ground as possible. These deployers will need to conduct impact assessments on deployed AI systems annually and redo the process within 90 days of any major changes.
For developers, the rules are a little more complex. Developers must ensure a general statement is available to deployers and other developers, which consolidates “reasonably foreseeable uses and known harmful or inappropriate uses of the high-risk artificial intelligence system.” In addition to this, in adherence to the Colorado AI regulations, developers must also present a document disclosing details such as the type of data that was used to train the high-risk AI systems, any limitations of the systems, the purpose of the AI, the benefits and useless of the system, and all the necessary information that the deployer will require in order to comply with regulations.
Developers also need to maintain documentation that covers their training data governance measures, performance evaluation metrics used, intended outputs, measures taken to mitigate risks, and guidelines surrounding the use of high-risk AI systems. Developers have to update their statements within 90 days of any modifications to the high-risk AI systems, to reflect the changes and their impact. Colorado’s AI discrimination bill foresees that there might be issues that cannot be predicted today, but strict documentation of every aspect of the AI’s functioning should make it easier to track and correct when problems do come up.
What Do Colorado’s AI Regulations Mean for Consumers?
Colorado’s AI discrimination law aims to protect consumers from any discrimination that might occur through unregulated AI systems that are put in charge of decision-making. Additionally, consumers can be informed about the systems being used to collect their data and how this is being used The users are allowed to opt out of systems they do not wish to participate in. In specific cases, users can also appeal decisions made by these high-risk AI systems and request human review. These kinds of safeguards are going to be critical in an AI-powered future.
Just to summarize everything we know, the main requirements of the Colorado AI bias law include:
- Protecting customers from discrimination
- Creating policies that actively predict potential risks of the AI systems
- Being transparent about every element of the AI functioning for those who find reassurance from the data
- Publicly disclosing the data that is being collected and what it is being used for
- Conducting impact assessments to track the AI performance over time and monitoring changes made to the systems
- Correcting any data that was wrongly processed by the AI tool
Colorado’s AI discrimination law is a big step forward for the state as it attempts to regulate and monitor the changes that are inevitable with the introduction of AI. It’s also true that the guidelines of the Colorado AI regulation are far from perfect. They do not apply to small companies that employ less than 50 full-time employees and they do not use their own data to train high-risk artificial intelligence systems. This does create room for some deployers to circumvent the law, but overall, Colorado’s AI discrimination law does appear to be a well-intentioned move toward the future.