Modeling Study Shows Weak AI Rules are Likely to Backfire
New research suggests that poorly designed artificial intelligence regulations could have unintended consequences, potentially making AI products less safe than if no rules existed at all. With comprehensive federal oversight of AI still lacking in the U.S., state governments have introduced their measures to address safety concerns. However, experts say it remains uncertain how these varying requirements influence the decisions companies make when developing and deploying AI systems. To explore the issue, researchers from Carnegie Mellon University and Cornell University created a theoretical model examining how different regulatory approaches affect developers of general-purpose AI systems and businesses that build applications using the systems. Their findings, published…