“The rise of AI” is not just a trend, but a significant shift that has profoundly influenced the lives of most people in the past 5 years. While it brings numerous benefits, the rapid evolution of AI also presents real and immediate risks and concerns.
While some fear that AI will eliminate millions of jobs worldwide, perhaps the more pressing concern involves consumer protection and data privacy. Machine learning requires data; the more meaningful the data, the better the outcomes. This data has to come from somewhere. This data has to be securely archived somewhere. This data likely passes through many virtual hands.
How is this retrieval, utilization, transfer, and storage of data being regulated? How are consumers protected against unwanted AI interactions and outcomes?
AI data and consumer protection policies are now emerging but are still in their infancy.
Here is something else to consider…
With the above concerns in mind, we must also ask, “how is consumer data interacted with?”
Imagine a scenario where AI machine learning algorithms, in their quest for efficiency, develop discriminatory tendencies– tendencies that unfairly and significantly impact consumers. This is not a far-fetched possibility, but a likely reality we must be prepared for. What happens when these tendencies become widespread?
This consideration becomes even more important when considering high-risk AI systems or systems that play a substantial role in making a consequential decision. Such systems include those that affect the provision or denial to any consumer of:
- Education enrollment
- Employment
- Financial opportunity
- Government service
- Healthcare service
- Housing
- Insurance
- Legal service
As AI technology becomes more integrated into the human decision-making process, “algorithmic discrimination” within any high-risk AI system can harm consumers.
Introducing: A New Groundbreaking Legislation
On May 17, 2024, Colorado enacted what is considered the first comprehensive AI legislation the nation has seen and will go into effect in February of 2026.
In short, this act creates duties and obligations for developers and deployers to use “reasonable care” to protect consumers from any “known or reasonably foreseeable risks” of “algorithmic discrimination” arising from the use of high-risk AI systems.
This act also requires that AI technology developers and deployers are forthright about the high-risk systems consumers will interact with. Impact assessments, risk management protocols, and a means to resolve discriminatory tendencies that arise are also a part of this act (you can find more detailed information here).
What Are the Expected Results and Implications of this Legislation?
As the first full act of its kind, we must recognize its significance.
An Existing Template
Other states will likely soon adopt similar legislation. We are already seeing the beginnings of this. It’s also fairly certain that legislation will continue to arise and evolve. Consumer protection will continue to be an important issue for lawmakers, especially in the wake of emerging AI.
Employer Implications
As it turns out, some of the most widespread high-risk AI systems currently on the market are in the talent acquisition and hiring space. Thus, discrimination that arises within these AI technologies could possibly introduce more liability to employers. Now that laws are emerging to protect consumers from AI flaws, there are legitimate avenues for consumers to seek damages. This introduces even more compliance responsibility into HR departments, particularly those that will lean heavily on AI.
Conclusion
This Colorado legislation is a signal of things to come. Employers will have to be even more aware of the technologies they use and how they may impact the people with whom they interact. Whether hiring discrimination originates from a human or a machine, it is discrimination nonetheless, and AI algorithmic discrimination is now becoming monitored and regulated.
As new laws and regulations arise, understanding and tracking them is becoming increasingly important. Understanding your current and future AI machine-learning systems and their potential vulnerabilities is also important.
You can find our archive of monthly regulatory updates here.
