Understanding Regulations for AI in the USA, the EU, and Around the World

lications facing rigorous requirements. This includes sectors like healthcare and finance, where AI can significantly impact human lives. The Act's broad nature means it applies across all sectors, while its focus on fundamental rights introduces new regulatory concepts. For example, HR software using AI is classified as high-risk due to its potential to impact livelihoods. Nick Reiners, a senior analyst at Eurasia Group, emphasized the importance of these regulations, stating, "We don't want to wait until a catastrophe happens before regulating."

Global AI Governance: US and China

While the EU leads with its comprehensive AI Act, other global players are also shaping their AI policy. In the US, AI regulation is gaining traction at the state level, with active debates in Congress. Nick Reiners noted, "AI regulation has become more likely than perhaps it seemed just a few months ago." Meanwhile, China has implemented sector-specific regulations, focusing heavily on consumer-facing applications. Shalini Kurupati, CEO of Clearbox AI, highlighted the geopolitical differences, stating, "In China, it's more about the development of the country rather than the citizens themselves." This divergence highlights the varied global approaches to AI governance.

Compliance: An Interdisciplinary Effort

Achieving compliance with AI regulations requires collaboration across various disciplines within an organization. Technical teams, including data scientists and AI engineers, must work closely with legal departments to understand the complex AI policy. Shalini Kurupati noted that even though the AI Act is a legal text, "Data scientists and machine learning engineers will have a lot of things to do, especially in terms of conformity assessment and technical documentation." Companies must ensure data governance, fairness, and transparency in their AI models and be proactive in their compliance efforts. This interdisciplinary approach is essential for meeting regulatory requirements and mitigating risks.

Data Governance and Fairness in AI

Data governance and fairness are central to AI compliance, as regulations like the EU AI Act emphasize the importance of transparent and fair data practices. Organizations must ensure that their data sources are compliant with existing data protection regulations, such as GDPR, and maintain high-quality data standards. Shalini Kurupati stressed the need for transparency, saying, "You have to prove that your data is of good quality and you're able to achieve fairness, representativeness, completeness of your data." Utilizing synthetic data can help address privacy concerns while ensuring secure and fair AI models. Companies must document their processes and be prepared for audits to demonstrate compliance.