FEATURE
The EU AI Act is expected to take effect next year, providing a unified framework for AI regulation across its 27 member states. However, AI compliance outside the EU remains a complex challenge. The UK has opted for a more flexible, pro-innovation approach, while the US continues to regulate AI through sector-specific and state-level laws. The result is that requirements for companies are constantly evolving.
History offers a lesson in how to reach standardisation. In the 19th century, Britain’ s railway system initially lacked standardisation, with different companies using incompatible track widths. This inefficiency led to the government’ s intervention, mandating a standard gauge that later became widely adopted. AI governance today faces a similar challenge – without a unified framework, businesses must navigate inconsistent compliance requirements across jurisdictions.
Right now, of those companies that are complying with regulations, these only meet local standards, just as the railway companies did in the 19th century. It’ s easy for them to believe that these are the only rules that apply, since there is little enforcement in place to stop them and disjointed guidance around compliance.
Setting guardrails
So, what can they do to future-proof their global AI projects? As AI grows more sophisticated, its capacity to generate impactful – but potentially risky – outputs will escalate. Companies must make a strategic shift and take responsibility for implementing their own AI governance frameworks and guardrails.
STANDARDISING AI OVERSIGHT DOES MORE THAN TICK A REGULATORY BOX. www. intelligentcxo. com
29