THE LONG VIEW

Bottom Line Up Front: The European Union’s proposal to amend the AI Act less than two years after its enactment—via the Digital Omnibus on AI—reflects a legitimacy crisis driven by rapid technological and geopolitical pressures, risking regulatory instability and undermining long-term legal coherence (Casey & Colonna, arXiv). Threat Identification: The core threat is regulatory whiplash—the accelerated revision of foundational AI legislation before implementation, driven by three interrelated dynamics: (1) the race for AI regulation, where the EU seeks to maintain global leadership; (2) the race for AI dominance, involving economic and strategic competition with the US and China; and (3) the race for regulatory connection, aligning with international partners to ensure market access and interoperability (Casey & Colonna, arXiv). Probability Assessment: High probability of further amendments to the AI Act within 2026–2027. The political momentum behind the Digital Omnibus on AI suggests legislative action is likely by late 2026 or early 2027, given the EU’s focus on competitiveness and regulatory simplification (Casey & Colonna, arXiv). Impact Analysis: The impact includes increased uncertainty for AI developers and deployers, higher compliance costs, potential fragmentation in enforcement, and weakened public trust in regulatory durability. Prioritizing political and operational rationalities over legal and cultural ones may erode the perceived legitimacy of AI governance frameworks across member states (Casey & Colonna, arXiv). Recommended Actions: (1) Establish a sunset clause with built-in review mechanisms in future AI legislation to accommodate change without abrupt overhauls; (2) strengthen stakeholder engagement to preserve legal and cultural rationalities; (3) develop interoperability protocols with key trading partners to reduce regulatory divergence; and (4) create a standing EU AI Observatory to monitor technological shifts and recommend evidence-based adjustments. Confidence Matrix: - Threat Identification: High confidence (supported by explicit framework in source) - Probability Assessment: Medium-High confidence (based on current political trajectory, though subject to legislative delays) - Impact Analysis: High confidence (logical extension of cited legitimacy concerns) - Recommended Actions: Medium confidence (inferred from analysis, not explicitly detailed in source) Citation: Casey, D., & Colonna, L. (2026). The Digital Omnibus on AI, Legislative Legitimacy and the Dynamics of AI Regulation. arXiv:XXXX.XXXXX [cs.CY].