May 2024

10 Things You Need to Know about Colorado’s SB205 Consumer Protections for Artificial Intelligence

Colorado's Governor signed SB24-205 into law, providing consumer protections for artificial intelligence (AI) that will come into effect on February 1, 2026. The law requires developers and deployers of AI systems to demonstrate reasonable care to protect consumers from known or foreseeable risks of algorithmic discrimination. Deployers must implement a risk management policy, while developers must take reasonable care to protect consumers and provide deployers with documentation. The law defines algorithmic discrimination as unlawful differential treatment or impact against individuals or groups on the basis of certain characteristics. The law does not apply to certain high-risk systems approved or used in compliance with federal agency standards.

NIST AI RMF Generative AI Use Case Profiles

The National Institute of Standards and Technology (NIST) has released draft AI RMF Generative AI Profile guidelines to help organizations manage risks posed by generative AI. The framework provides a roadmap for managing GAI-related challenges throughout the AI lifecycle, including risks such as proliferation of dangerous content, data privacy and information security. Proactive measures provided by NIST include establishing clear policies and guidelines, identifying potential risks, implementing metrics and benchmarks to assess the effectiveness of risk mitigation strategies and developing mechanisms for monitoring, detecting and responding to GAI-related risks in real-time. The framework emphasizes the importance of tailoring risk management strategies according to an organization's unique context and situation, prioritizing AI governance.

AI and IP at the Crossroads: How does the EU AI Act Approach Copyright Law?

The development of artificial intelligence (AI) technologies relies heavily on data. However, the use of copyrighted materials in training and testing AI systems may raise copyright infringement concerns. The EU AI Act introduces no exceptions to current copyright law but imposes obligations on providers of general-purpose AI models to comply with copyright protection. These obligations include establishing copyright policies and providing detailed summaries of training data used in the development of the AI model. The AI Office is tasked with developing guidelines and monitoring compliance with copyright obligations. EU copyright law provides several exceptions to copyright protection, including exemptions for text and data mining activities. Still, non-compliant operators may face harsh penalties under the Act and copyright claims under EU copyright law.

Why are copyright laws relevant to AI?

The emergence of Generative AI (GenAI) has raised numerous legal questions, particularly in the realm of copyright law. One of the central concerns surrounding GenAI technology pertains to its utilization of copyrighted material for training purposes, leading to a plethora of allegations of copyright infringement and lawsuits. Several countries have laws that could potentially permit the use of copyrighted materials to train AI, but there is a clear divide between AI developers and copyright owners on the issue. The US has proposed legislation addressing GenAI and copyright issues, while legal battles have also taken place outside the US providing insight into the global reaction to the issue of copyright and AI. The lawsuits surrounding Generative AI technology address several critical questions at the intersection of copyright law and AI development, including whether training a model on copyrighted material necessitates a license, whether generative AI output infringes on the copyright of the materials on which the model was trained, and the issue of liability for copyright infringement stemming from generative AI. Organizations need to be aware of existing laws and the rising number of legislations targeting copyright to ensure compliance.

Colorado passes law enacting consumer protections for AI

Colorado has passed a law to protect consumers in their interactions with AI systems, SB24-205, which mandates regulation of high-risk AI systems and prevents algorithmic discrimination. Developers of high-risk systems must take reasonable precautions to prevent algorithmic discrimination, such as disclosing information and conducting impact assessments. High-risk system deployers must implement risk management protocols, conduct impact assessments, and provide consumers opportunities to correct incorrect data. The law applies to any person who does business in Colorado, and enforcement lies solely in the power of the Attorney General. The law provides defenses for developers or deployers if they comply with a nationally or internationally recognized AI risk management framework. There are no specified penalties for violations, and compliance can be maximized with Holistic AI's Governance Platform.