Colorado lawmakers are preparing to revise one of the first comprehensive artificial intelligence laws in the United States, following months of tension between regulators, consumer advocates, and the technology industry.

A newly released policy framework outlines how the state may adjust its 2024 AI law before enforcement begins later this year.

At the center of the effort is a practical challenge: how to regulate AI systems that are already in use across hiring, housing, lending, and government services — without making them too difficult or costly to deploy.

Why the Law Is Being Revisited

When Colorado passed its AI law in 2024, it drew national attention for taking an early and comprehensive approach.

The law focused on “high-risk” AI systems, such as:

• Job applications
• Access to housing
• Financial decisions
• Government services

It introduced requirements aimed at preventing algorithmic discrimination and increasing accountability.

But soon after, companies raised concerns.  The requirements were too broad, too complex, and could significantly increase the cost of using AI systems. In response, the state delayed enforcement and formed a working group to revise the approach.

What the New Proposal Changes

The updated framework reflects an attempt to find a middle ground. Instead of imposing strict, one-sided responsibility, the proposal introduces a more shared model of accountability.

Developers would be required to:

• Disclose how their systems work
• Provide information about data sources and limitations

Organizations would be expected to:

• Inform individuals when AI is being used in decisions
• Use clear, plain language when doing so

This is a shift from the original structure, where responsibility was more concentrated in one place.

A New Approach to Liability

One of the most important changes involves liability.

The original law raised concerns because it could place a large share of responsibility on a single party, even when multiple actors were involved in how an AI system was developed and used.

The revised framework takes a different approach.

Responsibility would now be assigned based on who did what:

• Developers would be accountable for how systems are built
• Deployers would be accountable for how they are used

This reflects a more realistic view of how AI operates in practice.

The Ongoing Debate: Protection vs. Practicality

Even with these revisions, the outcome is not certain. Some lawmakers have already indicated that the proposal is only a starting point, and further changes are likely as it moves through the legislative process.

The broader tension remains.

On one side:

• Consumer protection
• Preventing discrimination
• Increasing transparency

On the other:

• Cost of compliance
• Impact on innovation
• Practical ability to deploy AI systems

Colorado is now trying to balance both.

What This Means in Practice

For organizations, the takeaway is less about one specific law and more about direction. AI systems are increasingly being treated like other regulated business processes. That means expectations around:

• Disclosure
• Documentation
• Accountability
• Oversight

are becoming part of how these systems are evaluated. At the same time, regulators are still working out how to apply those expectations in a way that remains workable.

The post Colorado Moves to Revise Its Landmark AI Law After Industry Pushback appeared first on Centraleyes.

*** This is a Security Bloggers Network syndicated blog from Centraleyes authored by Rebecca Kappel. Read the original post at: https://www.centraleyes.com/colorado-moves-to-revise-its-landmark-ai-law-after-industry-pushback/