California SB 53 and SB 942: How the US and Canadian AI Transparency Rules Apply to Your Organisation
In response to growing concerns about accountability in automated decision-making (ADM), California has introduced two bills that will soon set transparency standards for AI systems. While these rules primarily target US-based vendors, it’s crucial for Canadian organisations to understand their implications and prepare for similar regulations in the future.
SB 53 requires vendors of AI-powered products to disclose information regarding their ADM capabilities, ensuring users can make informed decisions about the technology they use. Similarly, SB 942 demands transparency around ADM algorithms used by healthcare providers. Both bills focus on the importance of understanding how these systems operate and making those insights accessible to customers.
These developments have far-reaching implications for organisations that buy or deploy AI in regulated sectors like healthcare, insurance, and financial services – primarily in Canada and the UK. They expose your organisation to potential risks, including:
1. Capability risk: Does the technology do what was claimed? If it fails to deliver on its promises, you may waste resources and face reputational damage.
2. Governance risk: Can the decision be defended when boards or regulators ask? Inadequate explanations of ADM processes can lead to regulatory non-compliance and potential fines.
3. Liability risk: What is your organisation accountable for when the AI errs? As the user, you’ll be held responsible if AI decisions result in harm.
To mitigate these risks, prudent organisations take immediate action:
1. Review existing AI solutions for compliance with new transparency requirements.
2. Assess ADM capabilities and understand how they operate within your business processes.
3. Develop strategies to address potential regulatory and governance issues.
Organisations may struggle to address these challenges independently due to vendor information asymmetry and internal teams being too close to the decision-making process. Without a neutral, independent perspective, it’s easy for organisations to overlook critical ADM elements or overestimate system capabilities.
Straven & Co examines AI decisions before they are acted on: stravenandco.com