US state legislators driving AI regulation despite resistance from industry chiefs

According to reports, Texas, Connecticut and Colorado are among states committing to tackling inherent bias within AI at a regulatory level, with measures designed to address issues across housing, healthcare and other key sectors in American society.

The drive hinges on a desire for enhanced accountability and transparency concerning how AI technologies shape individual citizens’ lives. Connecticut and Colorado are two states in which companies would be obliged to conduct and unveil “impact assessments” of their AI frameworks.

These evaluations would provide better understanding of the potential for bias within AI decision-making pathways, break down harvested date, AI’s role in decision-making, and the steps taken to offset risk of bias.

Despite their altruistic motives, these proposed measures have sparked a clash between civil rights factions pushing for increased supervision and industry players cautious of operational constraints and legal weak points.

The primary issue centres upon the processes that determine and guarantee accountability. While the new laws would hold companies more liable by obliging them to report any identified bias to the attorney general, detractors say that the self-reporting method could dilute the regulation’s strength, enabling discriminatory practices to continue.

David Edmonson, speaking on behalf of pro-tech group, TechNet, has advocated a measured approach to AI regulation that addresses risks while fostering innovation. However, concerns voiced by labour unions and consumer representatives highlight fears of companies taking advantage of legal loopholes, in moves that would eventually be damaging to workers and consumers.

The conversation drifts into the topic of legal recourse for victims of AI bias, with current proposals mainly limiting the ability to file lawsuits to state attorney generals, thus reducing individual citizens’ access to justice.

Underlining the pioneering nature of the policy efforts, Colorado’s Democratic Senate Majority Leader Robert Rodriguez, said:

“Every bill we run is going to end the world as we know it. That’s a common thread you hear when you run policies. We’re here with a policy that’s not been done anywhere to the extent that we’ve done it, and it’s a glass ceiling we’re breaking trying to do good policy.”

Know the risks

As deliberations continue, the outcomes of these legislative drives are poised to establish crucial precedents for the future of AI regulation, the end goal being to strike a balance between technological advancement and the protection of civil rights, ensuring fair treatment for all.

The issues fall into focus next month at PrivSec & GRC Connect Atlanta, where industry experts will discuss AI regulation and the challenges organisations face as they move to use emerging technologies ethically and compliantly.