After 12 hours of discussions, the EU and Parliament could not reach an agreement on the modifications to the AI Act, delaying the deal until next month.

      The breakdown of Tuesday’s trilogue reveals significant divisions regarding the exemption of high-risk AI systems incorporated in consumer products from the world’s strictest AI regulations. After 12 hours of discussions, EU member states and lawmakers from the European Parliament were unable to finalize a deal on proposed modifications to the landmark AI Act. As reported by Reuters, negotiations will continue in May.

      “A consensus with the European Parliament was not achievable,” stated a Cypriot official, with Cyprus currently holding the rotating presidency of the EU Council. This unsuccessful session marked the last scheduled political trilogue on the AI Omnibus, which includes amendments to the AI Act that is set to come into effect in August 2024, along with proposed adjustments to the GDPR, the e-Privacy Directive, and the Data Act. The Omnibus is positioned as a competitiveness measure intended to alleviate regulatory burdens on businesses, allowing European companies to compete more effectively with their US and Asian counterparts. Critics, including a broad coalition of privacy and civil rights groups, contend that it represents a rollback of hard-won protections disguised as simplification.

      The central unresolved issue from Tuesday’s talks was whether high-risk AI systems integrated into products already governed by EU product safety legislation—such as medical devices, toys, connected cars, and industrial machinery—should be exempt from the additional requirements of the AI Act. The European Parliament, with support from industry groups, has advocated for these systems to be regulated solely under their existing sector-specific laws. Conversely, the Council, representing member states, has been hesitant to agree to such a wide-ranging exemption.

      The Omnibus has faced ongoing criticism from researchers and civil society organizations, who argue that diluting the AI Act before its key provisions are even implemented risks undermining one of Europe’s most notable regulatory frameworks. Michael McNamara, the Parliament’s chief negotiator on the AI Omnibus, expressed in an interview with Tech Policy Press that while managing overlapping regulations can be complex, moving AI governance to sector-specific laws could ultimately lead to “deregulation rather than simplification.”

      Civil society groups have been even more explicit. Over 40 organizations signed a letter to the Parliament in mid-April, arguing that the proposed amendments undermine the fundamental rights protections of the AI Act, particularly concerning biometric identification systems, AI utilized in educational settings, and medical AI. The AI Act was widely regarded as a global benchmark when it was enacted.

      The urgency of the negotiations stems from structural factors. The key obligations of the AI Act for high-risk AI systems are scheduled to take effect from August 2, 2026, which is only three months away. The primary aim of the AI Omnibus is to push back that deadline to December 2, 2027, for stand-alone high-risk systems, and to August 2, 2028, for those integrated into regulated products.

      For this postponement to be legally actionable before the August deadline, a final political consensus, a formal vote in Parliament, endorsement from the Council, and publication in the Official Journal must all occur within a matter of weeks.

      If negotiations remain stagnant in May and no agreement is reached by June, the original August 2026 deadline will be upheld. This scenario would leave companies that rely on the extended timelines of the Omnibus suddenly facing immediate compliance requirements for which they may not be adequately prepared—an outcome that Brussels is striving to prevent.

      Additionally, the Omnibus includes one widely endorsed provision: a ban on AI systems producing non-consensual intimate images, including materials related to child sexual abuse. This measure was added to the package following the controversy surrounding nudification capabilities of Elon Musk’s Grok chatbot in late 2025, with both the Parliament and Council having aligned on it. The collapse of talks despite this consensus highlights how persistently challenging the issue of sectoral exemptions remains.

      The continuation of discussions in the coming month will determine whether the EU can assert that it is handling this process methodically, or if the world’s most ambitious AI regulations stumble just as their most stringent rules are expected to take effect.